In the traditional taxonomic system, as used since Linnaeus (1758) by the overwhelming majority of zoologists, the taxonomic information is indexed thanks to a system of Latin scientific names or nomina (Dubois, 2000). This system is organized in a hierarchical way by means of nomenclatural ranks successively fit into each other, from kingdom and phylum to variety and form. The International Code of Zoological Nomenclature (Anonymous [ICZN], 1999) (“the Code” below) provides Rules for the unambiguous and automatic allocation of a unique and universal valid nomen for each taxon, at least for most nomina (excepting the higher and lower ranks). From a nomenclatural point of view, a species is simply a taxon of rank lower than genus (or subgenus or species-group if relevant) and higher than subspecies (or subspecies-group if relevant). Species as a nomenclatural rank is essential outside systematics. For this purpose, which is intended mostly to non-systematists and even to non-biologists, it is indispensable to have a universality of use over the whole zoology. All taxa of rank species must be designated in the same way in the whole world, by a Latin binomen such as Drosophila melanogaster.

Several modes of formation of gametes have been described in animals, some involving a normal or modified meiosis (with a reductional cell division), and some an ameiosis (consisting only of normal mitoses). For reasons given above, the term gametopoiesis is here used instead of the traditional term “gametogenesis”. The term metameiosis (Dubois, 2008a) designates any modified meiosis, i.e., any meiosis different from eumeiosis (Battaglia, 1945 and Battaglia, 1955) as described in all treatises of biology. The term metameiosis is in no way descriptive but simply points to the presence of an “abnormal meiosis”, irrespective of its peculiarities.

The study of all the “strange species” mentioned above suggests that most, if not all, cases of natural gametopoiesis in animals may be referred, from the viewpoint of their results in terms of genetic complement of the gametes, to five main categories, very briefly characterised below but presented in more details elsewhere (Dubois, 2008a and Dubois, 2009a). What is relevant and important here, in the frame of a reflection on the taxonomy of these entities, is not the cytological processes involved in the formation of gametes, but the genetic composition of the latter.

Mixopoiesis is the kind of gametopoiesis at work in most mayrons, but also in some kyons. It involves an eumeiosis.

In phylactopoiesis, the gametes produced are all diploid and contain a genome identical to that of the mother. This may be the result of two distinct gametopoietic pathways: ameiosis (in apomictic and some gynogenetic entities) or metameiosis (in some gynogenetic entities) (Dawley and Bogart, 1989 and Simon et al., 2003).

Airetopoiesis concerns the taxa with automictic reproduction, the metameiosis of which involves an intracellular doubling of the parental chromosomal stocks, followed by a meiotic reduction and then by a fusion that produces diploid gametes. This results in both heterozygous and homozygous gametes, but in the long run the entity tends towards becoming homozygous (Simon et al., 2003).

Elasopoiesis occurs in some taxa of hybrid origins, the kleptons. Their metameiosis produces only “pure” gametes containing only chromosomes from one of the two parental mayrons.

Finally, in the recently discovered tychopoiesis (Bogart et al., 2007), the gametes produced contain one or several hemigenomes of variable origins, either maternal or paternal. These gametes depend on sperm for the initiation of their development, but may, or may not, be fertilized, so that a great variability is observed in the offspring of any given female.

Several modes of initiation of embryonic development in an ovum or kinetogenesis (Dubois, 2009b) exist, some only being known to occur spontaneously in nature.

Zygogenesis (Dubois, 1991, Pilger, 1989, Pilger, 1997 and Tomlinson, 1968) is the fertilization of an ovum by a sperm, producing a zygote. In this system, except for identical twins, each individual possesses its own assortment of chromosomes and genes.

Parthenogenesis is the development (either spontaneous or experimentally induced) of a virgin ovum without any contact with sperm. Apomixis concerns ova of phylactopoietic origin, whereas automixis occurs in ova of airetopoietic origin. The offspring of a parthenogenetic female may thus bear the same genetic complement as the mother, or not.

In gynogenesis (Wilson, 1925), the development of the ovum is initiated (either spontaneously or experimentally induced) by its contact with a sperm, but without incorporation of the genetic material of the latter into its nucleus. All the children of a female thus have the same genome.

In androgenesis (Verworn, 1891 and Wilson, 1925), the development takes place only with the paternal chromosomes, after expulsion or destruction of the maternal chromosomes. Although rare, this phenomenon exists in nature in some animal groups (McKone and Halpern, 2003).

Two modes of kinetogenesis recently described can be designated as cases of pseudozygogenesis (Dubois, 2008a). In these cases, a pseudozygote is obtained by fusion of two gametes from one or several individuals of the same sex. In corydogenesis, the fusion concerns two sperms introduced in an enucleated ovum, whereas in lesbogenesis the fusion is between two ova.

The specion concepts that we considered above (mayron, klonon and klepton) underline the patterns of genetic flow within the entities examined or between them. This approach is not sufficient. It describes specions as static, fixed entities, and ignores the conditions of their appearance and evolution. In an evolutionary perspective, it must be completed by a dynamic approach of specions as historical entities. Here we will consider the patterns of speciation, i.e., the characteristics of the “birth” of a specion and its relationships with its ancestral specion(s). In the discussion that follows, speciation is considered, by definition, as a phenomenon unique and irreversible, separating in a definitive way (but sometimes incompletely, in the cases of kleptons) two or more evolutionary entities.

Because its purpose is taxonomic, the following discussion is strictly targeted. It is not focused on the modes but on the patterns of speciation. The speciation modes, i.e., the mechanisms operating in the separation of entities, can be classified according to the geographic relationhips between the latter (allopatry, sympatry, etc.) or to the biological mechanisms (genetic, behavioural, etc.) involved. As concerns speciation patterns, they simply refer here to the structure of the phylogenetic relationships and of genetic transmission between the individuals and the lineages.

The patterns of speciation in the animal kingdom can be referred to three main categories (Dubois, 2009a). In monogeny, a specion is replaced in nature by another one; in diplogeny, a specion gives birth to two distinct specions; and in mixogeny, a new specion is produced through hybridization between two, or more, specions. These three patterns of speciation were reviewed in detail elsewhere (Dubois, 2009a) and are only briefly presented below.

Table 1 and Table 2 give the main evolutionary, genetic and reproductive characteristics of the eidonomic categories (categories of specions) presented above.

The main category, to which is most generally applied the “species concept”, is the “mixiological species”, or BSC, here termed mayron. Mayrons are usually panmictic bisexual entities, with complete eumeiosis and ovum fertilization, and therefore a complete reproductive independence from all other specions and a non-clonal heredity with recombination between parental genomes at each generation. Except for identical twins, each individual of such an entity possesses its own genetic identity, no two individuals having the same genome. It is not necessary that all the individuals of a mayron have the reproductive mode summarised above for the taxon to be referred to this taxonomic category. This reproductive mode may be “dormant” for one or a few generations or absent in some individuals without challenging the fact that the entity in which they belong is bisexual and (on the whole, and sometimes virtually) panmictic. For example, the taxa with facultative or cyclical parthenogenesis include individuals that may reproduce without ovum fertilization, but these events are not at the origin of real clones, as their descendents, sooner or later, take part again in bisexual reproduction and therefore in a redistribution of chromosomes and genes. Even in very particular cases like some all-triploid taxa mentioned above, the existence of metameiosis in some individuals is associated with eumeiosis in other individuals, so that, in the long run, genetic mixing occurs. These entities, in which panmixy exists only over several generations, should be referred to the taxonomic category of mayron, although in the subcategory of heteromayron.

Although several other eidonomic categories were recognized above (klonon, with two subcategories, and klepton, with three subcategories), the main opposition is between mayron and all the other categories, which may be united under the common designation of kyon. Specions of this second group are probably much less numerous in nature than those of the first one. They show very diverse characteristics which challenge generalisation. They will most likely have to be complemented by a few other very rare or still not understood situations which will probably deserve the creation of other eidonomic categories – in particular, entities with androgenetic and corydogenetic kinetogeneses (see above). According to the information currently available (Loxdale and Lushai, 2003), these entities are all of hybrid origin, but they all persist over long periods in nature, contrary to “normal hybrids”.

As a matter of fact, and contrary to what was long believed by evolutionary biologists, all these “particular cases” are not “evolutionary dead ends” bound to rapid extinction. Despite the fact that some of them depend on sympatric mayrons for their reproduction, they may perpetuate themselves over millions of generations (Bi et al., 2009 and Bogart et al., 2007). In some particular conditions (e.g., small isolated populations of klonons of lizards and snakes on very small islands), kyons are advantaged compared to bisexual mayrons. Some kleptons clearly show heterosis compared to their parental specions. However, they may not replace them completely in nature because they need them for their reproduction and perpetuation, so that an equilibrium must be found that allows the maintenance of both taxa in sympatry, except in some cases involving polyploidy where some members of the synkleptic complex can provide sperm and allow elimination of the parental species from some populations (Bogart et al., 2007, Polls Pelaz, 1991 and Polls Pelaz, 1994). Moreover, in some cases, it is likely that kyons can be at the origin of allopolyploid mayrons, thus so to speak “restoring” a “normal” situation after an intermediate “abnormal” stage (Dubois, 1977). All these questions, and many others (concerning behaviour, ecology, competition, selection, etc.) are of great interest for evolutionary biologists, so that, although unusual, these “strange species” deserve close attention.

As for mayrons, although they are the “normal” situation for many animals, this should not obscure the fact that intermediate and temporary situation are not uncommon. Categories like submayron, quasimayron, vicemayron and promayron testify to the fact that specions are not eternal “essences”, corresponding to a “type” and to characters given once and for all, but living, evolutionary entities, the fate of which is not written in advance but will depend on the conditions of environment, of competition and other interspecific relationships. In other words they are changing entities in the permanent dynamics of evolution.

These taxonomic categories are no doubt simplifications of the real evolutionary situations, which all have their own particularities, but they provide a rather rich and diversified frame of reference for such situations, which allows to account rather finely for their main features.

The approach presented here is quite different from that adopted by many contemporaneous authors who showed interest in the “species problem”. Most of them were clearly obsessed by the search at all costs of a “unified species concept”, allowing one to account altogether for all situations observed in nature. This search seems to take its roots in philosophical convictions tied to a reductionist conception of science, according to which, as stated by Aristotle, science only deals with universal concepts. This formula applies to non-historical sciences like physics or mathematics but not to the study of organic evolution. There exists no theoretical reason why all animal organisms in the world should be referred to a single kind of evolutionary unit or “brick of evolution”. The analysis presented above shows that nature is richer than our a priori expectations. To account for the observed facts, it is necessary to recognize several eidonomic categories of “bricks of evolution” or specions. These major categories are alternative, that is, they are not reducible or infeodated to each other. These different evolutionary situations (and most probably others) exist “in parallel” in nature, and the role of evolutionary biology and taxonomy is to identify them, through a fine analysis of their concrete particularities, rather than starting from general and reductionist “models”.

The requirement for a “unified species concept” results in a considerable impoverishment, in an abusive simplification of the situations observed. Given the diversity and richness of real evolutionary situations, and the incompleteness of the data we have on many organisms (in particular fossils), for which we lack information on gene flow and the dynamics of hybrid zones, the only “unified species concept” that could be implemented would be the “smallest common denominator” to all situations, and the only possible species concept would be that of simpson. This concept is misleading, as it would lead us in certain cases to treat as two “species” two groups of individuals geographically isolated, even if this isolation is of short duration and has entailed no evolutionary divergence between these two groups, which remain able to hybridize freely, with unrestricted bidirectional gene flow between them, if they come in contact again. Besides, the use of this unique species concept in all situations results in important information losses, for two distinct reasons.

The first reason is that this concept does not account for some evolutionary situations (the kleptons), where a well-individualised but partial lineage depends for its perpetuation of the intervention at each generation of a sympatric bisexual entity. This mayron would not be threatened with extinction if the klepton was to disappear. There exists therefore no mutual dependence between a mayron and its associated klepton. The mayron is indeed an “independent lineage”, but the klepton is a “half-lineage” that is dependent from the mayron. There is a formal way to solve this problem in order to try and “save” the “unified species concept”: it is simply to negate the existence of these peculiar entities, and to include formally each klepton into the mayron that permits its reproduction and perpetuation, despite the fact that the two forms are fully distinct genetically, cytogenetically, morphologically, ethologically, bioacoustically and ecologically, may have existed as such for millions years, and may have distributions covering a whole continent, where they play fully different roles in the ecosystems (Dubois, 1977, Dubois, 2008a, Dubois and Günther, 1982 and Graf and Polls Pelaz, 1989). For example, some authors (Frost and Hillis, 1990) considered that from a taxonomic point of view the entities that transmit a hemigenome from generation to generation in a hemiclonal way (zygokleptons) do not deserve to be treated as taxa of their own, just like the males of organisms bearing a heterochromosome Y do not belong in a specion different from that of the females having an XX genome! However, comparisons may be misleading, and this one is fully inadequate: males and females of all mayrons having a chromosomal sex determinism cannot breed and therefore subsist without each other, they are in a situation of mutual dependence! If our aim is to found eidonomy, the “science of species”, not on a priori general models and theories, but on observations of the real characteristics of natural entities, we need a peculiar category for those which depend, at each generation, on gametes provided by another, more “normal”, entity.

Ignoring generally the case of kleptons, the supporters of the simpson concept claim that it accounts for the evolutionary situation of klonons as well as of mayrons, both kinds of entities being qualified by them as “independent lineages”. There is however a difficulty here, as the meaning of the term “independent” is unclear in this context. It cannot mean the absence of etho-ecological interactions, as these are frequent between sympatric specions, especially when they have the same origin, and it must therefore mean that there are no genetic interactions between individuals belonging to distinct specions – which in return means that such interactions do exist within a specion. Then, if this is indeed the case for mayrons, it is not true for klonons. In the latter, the only genetic link is vertically from parent to offspring, but never horizontally between individuals of the same generation. In these conditions, what should be considered to constitute an “independent lineage”? Is this the group of all individuals having the same genome, before mutations have appeared? Such an interpretation would be a return to an essentialistic, typological conception of lineages and of taxonomy. If the specion is understood literally as an independent lineage, any distinct founder event must be interpreted as the origin of a new lineage, and therefore as a distinct taxon. From a purely formal point of view, every time a female gives birth by parthenogenesis to two daughters, at least one new lineage appears, independent from the other parallel lineage, even if it possesses exactly the same genome. Should we then recognize each of these lineages as a distinct specion? As it would doubtless appear absurd to recognize then two “species”, even the supporters of the “unified species concept” acknowledge that in this case the two “sub-lineages” must be considered to belong in the same taxon, but then they are obliged to modify their definition of “independent evolutionary lineage” and to write such vague statements as: “species-level lineages in asexual organisms, if they exist at all, must result from processes other than reproduction” (de Queiroz, 1998).

The solution to this problem requires one to take in consideration the founder event of the parthenogenetic lineage. Of course, a single founder event may be enough to start a klonon, but in cases of contact between two mayrons in a rather wide zone, it is not rare to observe several distinct events of hybridization. The different parthenogenetic lineages thus initiated are definitely isolated from each other from a genetic point of view. However, it would be fully irrelevant and impracticable to recognize each of them as a distinct taxon. Such a hyper-analytical approach would not bring an appropriate tool to all users of taxonomic data, many of whom are neither taxonomists, nor phylogenetists, nor geneticists. Although some “purists” of evolutionary or cladistic thinking tend sometimes to forget it, taxonomy is not meant only for themselves, but has among its main functions that of providing a tool to all biologists interested in biodiversity. This tool must be efficient and rigorous, but of clear and easy use. To recognize in a purely formal way such countless taxa, virtually identical in most respects, for the only sake of respecting an a priori theoretical decision, would not do a service to these users and would contribute to a bad image of taxonomy among biological sciences. The genomes of all individuals of the different parthenogenetic lineages resulting from distinct phenomena of hybridization between the same two parental mayrons result from different combinations of genomes originating from the same two initial gene pools. These individuals therefore share the same global genetic pool, and they have most similar behaviours, ecologies and habitats. Although they do not recombine their genomes to produce a common descent, they belong in a single evolutionary entity, which occupies a peculiar ecological niche in the ecosystem, and they must be referred to the same klonon.

A similar situation exists in viruses, in which genetic transmission is also clonal and all individuals genetically very similar, although not identical, because of the frequent occurrence of new mutations. The definition of the species category in virology adopted by the International Committee on Taxonomy of Viruses reads as follows: “a virus species is a polythetic class of viruses that constitute a replicating lineage and occupy a particular ecological niche” (Pringle, 1991, Van Regenmortel, 2007 and Van Regenmortel, 2010). This concept is equivalent to that of klonon as recognized here for parthenogenetic animals, although different in the details due to the many differences between viruses and metazoa.

The second reason why the generalisation of the use of the concept of simpson to all specions results in an important loss of information is that it discourages taxonomists to pay close attention to the fine particularities of the evolutionary situations surrounding speciation in the cases of hybrid zones. Whereas during the 20th century many evolutionary biologists had worked on the characterisation of these situations, to the elaboration of concepts and criteria allowing their detailed eidonomic treatment, the “all-cladistic” approach currently dominating invites taxonomists to abandon these fine analyses, to replace them by brutal and schematic affirmations such as “as soon as there are hybrids, we are inside the same species” (Samadi and Barberousse, 2006) or “as soon as two groups of individuals are geographically separated, they are distinct species” (de Queiroz, 1998). Such rigid and purely formal approaches are at variance with the whole tradition of zootaxonomy and “redefine” drastically the terms “hybrid” and “species” in a very confusing way. They do not encourage biologists to study the concrete features of the situations of speciation, although these constitute an irreplaceable mine of information on the reality of speciation and on evolution in progress (Dubois, 2008b). This reality is more complex and difficult to analyse, understand and translate in taxonomic terms than as suggested by these reductionist approaches. The taxonomic categories presented above, even if they are more complex than any “unified species concept”, remain nevertheless simplifications and generalisations, particularly as concerns all the situations here gathered in a partly artificial manner in the heterogenous category of kyon.

A final difficulty with the simpson concept as an “independent lineage” is that it does not provide any clue for distinguishing simpsons from all other more inclusive taxa or “clades”, which are also cladistically defined as “independent lineages”. The idea that the latter would be “multiple lineages” and the former “single lineages” (de Queiroz, 1998) does not hold, because no clear and operational definition of “multiple lineage” is available, as we have seen above in the case of klonons.

The concept of mayron doubtless corresponds to a real situation in nature, but concretely, in many cases, the data are insufficient to identify mayrons with certainty, and the operational difficulties of the concept are real. Although theoretically well supported, it poses real problems of implementation in many concrete situations, as it can be tested only within a unit of time and place. It is a strong stimulation for the deepening of works on the structure and dynamics of hybrid zones or on the study of the status of allopatric entities deriving from a single ancestral specion, for which it is often impossible to know whether in nature a reciprocal gene flow would establish if they came in contact again.

Two entities, which never breed together in nature can do it in captivity or in perturbed ecological conditions. The situation can then be clearly interpreted only in one case, when repeated crosses between the two entities have proved sterile: then, it can be quite safely concluded that the two entities are two distinct mayrons. But the reverse is not true, as this relational criterion (Dubois, 1988) or relacter (Dubois, 2004) is not symmetrical (Dubois, 1977, Dubois, 1988 and Dubois, 1998). If artificial crosses produce viable and fertile offspring, other criteria will have to be used, as the mayron concept is a naturalist concept, which applies and is meaningful only in nature. From the viewpoint of the mayron concept, it is only the dynamic reply in nature to the restoration of the contact between the two entities which allows a clear interpretation. What happens then in a contact zone is directly correlated neither with morphological resemblance nor with genetic similarity. The so-called “genetic distances”, either measured by the traditional methods of protein electrophoresis or through nucleic acid sequencing (including through “barcode” approaches), do not allow to predict what would occur in case of restored parapatry or sympatry. In all cases where we observe in nature isolated hybrids or hybrid zones between two entities, the relevant question is whether there exists a reciprocal gene flow between them or whether this gene flow is unidirectional, towards the hybrid zone where it “falls” like in a “black hole” from which no gene flow returns towards the two parental entities. In many cases, reliable replies to these questions are difficult to obtain for lack of data.

Because of the difficulties in the use of the mayron concept in allopatry, it is often necessary in such cases to do with a default solution. A usual practice (e.g., in “barcode” approaches) is to call on inference, i.e., on comparisons. If entities separated by a certain degree of divergence behave in sympatry as distinct mayrons (absence of significant gene flow between them), it is acceptable, in the absence of contradictory evidence, to consider that other entities in the same taxonomic group showing a similar degree of divergence are also distinct mayrons. Another solution may be through the “temporary” use (that in some cases may last long) of other concepts of specion, particularly that of simpson, which is of simple use although it has a limited biological meaning.

Such solutions are certainly not “ideal”. Should these difficulties lead us to reject the concept of mayron as “non-operational”? Certainly not. Science does not always progress in a straight line, from certainties to certainties. In many cases, leaving some questions open, instead of closing them artificially by declaring that they do not exist, may be an important stimulating factor for a deapening and improvement of research. The differentiated use of more or less “precise” concepts according to the available information is rather frequent in science and has nothing shocking. Let us consider another example. Nowadays, the phylogeny and higher taxonomy of vertebrates is more and more based on phylogenetic analyses relying on nucleic acid sequences, combined with studies bearing on large numbers of characters obtained from morphology, cytogenetics, etc. But when one turns to fossil vertebrates, almost none of these pieces of information is available: we cannot study their soft parts, their chromosomes, their behaviour or, except in rare cases, their nucleic acids, so that most of the taxonomy of these organisms must rely on information obtained from the study of their skeleton. In order to build a “unified taxonomy” of all vertebrates, including both fossil and living taxa, should we then use their “smallest common denominator”, that is, should we base all the systematics of vertebrates on the characters of their most often fossilised parts, their skeleton? This would be absurd. We use as much information as available in all cases, and when information is missing we do what we can with the available data.

The same applies to eidonomy. Like in all other scientific domains, the use of the “smallest common denominator” would entail a loss, sometimes gigantic, of information. In concrete science, eidonomists make use of as much information as possible, i.e., more or less according to the situation. It is sometimes possible, when reliable and detailed data are available on a bisexual entity, to use correctly the category of mayron. In other cases, data are wanting and this is impossible. It is then appropriate, at least in a provisional way, to use inference approximations, or to appeal to the category of simpson, even if this interpretation may later have to change when new data are obtained. The same applies when biologists discover in nature an entity having abnormal biological characteristics, such as being constituted only of females. This information, if it is reliable, shows that this entity cannot be a mayron, but does not tell us, by itself, if it is a klonon or a klepton. As long as complementary data are unavailable, it is advisable to treat this entity simply as a kyon.

The arguments presented above against the “unified species concept” apply only, let us stress it again, to the concept of “species as a taxonomic category” or specion, but not to the nomenclatural rank species. In order to play fully its function of universal communication about living organisms, usable both within biology and outside of it, it is indispensable that a single nomenclatural hierarchy be used for the whole animal kingdom, independently from the taxonomic theories adopted by the authors (Dubois, 2005, Dubois, 2007 and Dubois, 2008a). The nomenclatural rank species is crucial for the whole biology, and each living organism must be referable to a taxon at this rank, designated by a Latin binomen. This operation is distinct and independent from the allocation of this taxon to one of the taxonomic categories discussed above. In order for us to be able to communicate about them, all taxa must bear nomina. Kyons, just like mayrons, must be referred to by Latin binomina governed by the international nomenclatural rules of the Code, and not by any other system of numbers, letters, codes, abbreviations or plurinominal designations that are not real nomina (Avise, 2008, Bogart et al., 2007, Schultz, 1961, Schultz, 1966, Schultz, 1967 and Schultz, 1969). When what is at stake is simply to mention these taxa in official lists, in publications of physiology or in the medias, it is often enough to mention the two terms of the binomen, substantive and epithet. In contrast, in more specialised publications, intended for taxonomists or evolution biologists, it may be useful to indicate briefly, by a symbol placed between the generic and specific nomina, to which taxonomic category the taxon designated by this binomen belongs (see Table 1). It is thus possible to combine the indication of the nomenclatural rank species of these taxa (through their Latin binomen) and that of their fine eidonomic category (through the possible presence of symbols), the latter only providing information on the evolutionary peculiarities of the organisms and entities at stake.

In the light of all the facts presented above, the traditional representation of evolution traditionally credited to Darwin as a tree appears to be an inexact oversimplification. This is not only because this image carries a teleological connotation and that Darwin himself possibly favoured rather that of the coral (Bredekamp, 2003). This is mostly because some of the branches of this tree may, long after their separation, unite again to produce new branches. Even if it may appear shocking in regard of the current “dogma” of the “universal tree of life”, this phenomenon is not “exceptional” or “trivial” regarding the general pattern of evolution. It would be desirable to replace the image of the tree by that of the network. More than a tree, the branches of which always go irreversibly away from each other, the evolution of organisms reminds a network like those of some large tropical rivers, the arms of which split and then meet again, or the anastomoses of biological circulatory systems. This reticulate nature of biological evolution is a most general fact. Besides the phenomena of lateral gene transfer in prokaryotes, the importance of which starts only being understood (Doolittle, 1999), those of interpecific hybridization among complex organisms like metazoa are now known to be very widespread and to have played a major role in evolution, even in unexpected cases (Patterson et al., 2006). The image of the tree is misleading regarding evolution and has no operational universality regarding taxonomy (Doolittle, 1999 and Dubois, 2005), so it should be abandoned.