INFOWEB 11 : Costa's Catalog of aplocheiloid Killifishes of the World

From Jean H. Huber
Private address: 7 Bd Flandrin, 75116 Paris, France
M.N.H.N., Ichthyology, 43 rue Cuvier, 75231 PARIS Cedex 05.
e-mail :, [today both inactive]
S.F.I. : Société Francaise d'Ichtyologie (same address).

Paris, December 27. 2009 [updated on April 29. 2012, April 13. 2015 and July 1. 2015… Nimbapanchax].

Dear Colleague, dear Aquarist!

A new book by Wilson Costa, with a listing of all Aplocheiloid with fresh and provocative positions, leading to a major reappraisal of the generic assessment in Killi-Data online 

First of all it is important to mention that the book was only obtained for a short analysis at the end of 2009 but the book was already published in 2008, maybe as early as March [reference : Costa, W.J.E.M. 2008. Catalog of aplocheiloid Killifishes of the World. Universidade Federal do Rio de Janeiro Ed., Rio de Janeiro, Brasil, 127 pp. (dated, 3 March 2008)]. The reason why is that the book was poorly distributed by the Rio de Janeiro University, more as a personal list (as described by Wilson Costa himself) than as a scientific contribution and more as a way to express the author's views and differences on systematics.


Second the book looks apparently as a minor publication because it lists all Aplocheliloid species (Aplocheilidae, Rivulidae, Nothobranchiidae) with status, types, type locality and sometimes systematic comments on the basis of the Catalogue of Fishes (California Academy) or of Fishbase, while both Killi-Data or Killifish Master Idex provide with much more information… however it appears to be a false impression immediadiately after a quick reading and a look at the number of species that Costa considers as valid : 617 valid species, a number estimated much higher (by 20% ?) than reported in other systematic databases for that subgroup of Killifishes. The reason why is given by Costa at the end the book : he considers any name that he has not studied, or old and not studied by a professional scientist, as provisionally valid (hence he "revalidates" names that have been considered as synonyms since long, e.g., Aplocheilus affinis, andamanicus, dorsomarginatus, kuhlii, mattei, vittatus, or Epiplatys acuticaudatus, baudoni, marnoi, etc.) ; as a consequence, the Asiatic genus Aplocheilus which usually includes 5 or 6 species (Scheel, 1968), ends up with 14 possibly valid species ; this is more a personal suspicion towards previous synonymisations notably when amateur systematists are concerned (he states that "as a consequence of their great popularity among aquarists, taxonomy of killifishes is characterized by the wide contribution of amateur ichthyologists ; however, the theories of systematics are often unfamiliar to them, generating a series of mistakes when approaching topics as phylogeny, species concepts and biogeography", a position that is not shared by Killi-Data which promotes cooperation between researchers, professional and amateurs, and which do not make any difference between them).


Third the book becomes more significant when it tackles systematic concept such as the species concept and Costa's views are clear : he does not accept the subspecies unit (and erect all subspecies to the species level, since according to him "there is no sense in to recognize arbitrarily defined species subunits") ; as a consequence, subspecies reported by various authors as such, which are clearly diagnosed by distinctive features (e.g., Aplocheilus dayi dayi, Apl. dayi werneri) are considered as full species. Besides, he recognizes as valid species, names mentioned in inadvertent descriptions (e.g., today in Simpsonichthys : Cynolebias costae Baker and not Cynolebias costai Lazara, like for the case of Cynolebias antenori vs. heloplites), which is also a major change from general understanding (here, of the ICZN code).


Fourth the book becomes important when it goes to generic systematics because 48 genera are then considered as valid, a huge increase to the general understanding and acceptance ; basically all subgenera of west-african Nothobranchiinae are raised to the full genus rank ; for example, he moves Aphyosemion subgenera as genera (but he does not provides with a new diagnosis), e.g., for Raddaella, Chromaphyosemion, Mesoaphyosemion, Kathetys) ; he also creates, with very short descriptions, 6 subgenera of Austrolebias (Acantholebias, Acrolebias, Argolebias, Gymnolebias, Cypholebias) and 1 subgenus of Cynolebias (Bathylebias) ; Austrolebias luteoflammulatus is the type species of Acantholebias (diagnosed by a unique color pattern in male consisting in pale brown third anterior part of sides and silvery posterior 2 thirds part of sides, with 6-11 dark bars), Austrolebias carvalhoi is the type species of Acrolebias, Austrolebias nigripinnis is the type species of Argolebias (diagnosed by a dark gray Pectoral fins with bright blue iridescence in male), Austrolebias gymnoventris is the type species of Gymnolebias, Austrolebias robustus is the type species of Cypholebias, and Cynolebias griseus is the type species of Bathylebias (diagnosed by scales absent on venter, dark marks on suborbital and supraorbital regions missing in living specimens, dark gray anterior sides with lighter gray bars without brilliant colors in male) and in the end all these moves allow him to revalidate Megalebias as a subgenus (previously synonymized with Austrolebias).


In total this books appears a major milestone in our understanding, not of Killifishes, but of Costa's strategy and philosophy, and this is not minor because he is the researcher who by far has described the highest number of new names (species, genera) ever (and who also has collected so much in Brasil). There are according to our analysis at least 3 observations on Costa's approach :

  • all taxa previously considered as simple synonyms are to be restudied by him (or another professional scientist) and this fresh view can only be strongly supported by Killi-Data
  • all subspecies are to be considered as full species and this provocative view is to be considered on a case by case basis by Killi-Data (not as a general rule) : while it may be solid when variations are distinctive and populations are strictly allopatric (even if the 2 subspecies are breeding readily in aquarium… they do not play that game in nature), it may prove to be thorny when the geographic separation is not clear-cut and when the patterns show a continuous variation such as for Aplocheilus dayi dayi, Apl. dayi werneri (not to mention, when 2 subspecies have been described as such and both or one of them are simply unknown live, e.g. for Rivulus chucunaque chucunaque and Rivulus chucunaque sucubti
  • all subgenera which gather each a distinct morphospecies are to be considered as full genera and this conceptual view is to be considered as an opinion and not followed by Killi-Data, at least until a full solid rediagnosis is proposed ; for example, Mesoaphyosemion is not considered by Killi-Data as a distinct genus (molecular data have shown that its components are scattered over several morphospecies and a solid diagnosis has failed, yet, to be published, like for most Aphyosemion subgenera).


Coincidentally, at about the same time this book was at hand, Sonnenberg & Bush published [Sonnenberg, R. & E. Busch. 2009. Description of a new genus and two new species of killifish (Cyprinodontiformes: Nothobranchiidae) from West Africa, with a Discussion of the taxonomic Status of Aphyosemion maeseni Poll, 1941. Zootaxa, 2294: 9] a revision of Archiaphyosemion with currently 4 species and split it into 2 units with one species separated in Archiaphyosemion (guineense), as a monotypic genus and the others in a new genus Nimbapanchax. The new genus is not based on a strong diagnosis (lower vertebrae counts (vertebrae with pleural ribs 13-14, vertebrae with haemal spines 14-16, vs. resp. 14-15 and 16-17) but is evidenced by the molecular tree.

This strategy, very much in line with Costa's, is equivalent to name a genus for each morphospecies (and in the present case, the morphological differences between guineense and maeseni from the aquarium strain are thin). At the research level it is perfectly acknowledgeable, but for general usage in Killi-Data (and for aquarists) it bears 3 major weacknesses :

  1. it creates a lot of names with little added value (if any)
  2. it is more a question of philosophy to name all tree branches included within molecular trees (splitting vs. lumping) than of decisive results (for example, Parenti in 1981 changed in many ways the systematics of killifish with major osteological findings, but she described virtually no new genus name)
  3. it is not in line with ICZN code of nomenclature which supports stability and reasoned conservatism.

[addendum (April 13. 2015), : the analysis of a new work by Costa {ref. Costa, W.J.E.M. 2015a. Comparative Morphology, Phylogeny, and Classification of West African callopanchacine killifishes (Teleostei: Cyprinodontiformes: Nothobranchiidae). Zoological Journal of the Linnean Society} on genera Callopanchax, Archiaphyosemion, Scriptaphyosemion, Nimbapanchax shows that several new distinctive micro-osteological characters are added as diagnoses, however the diagnosis of Nimbapanchax is unfortunately based on a single, minor and weak character (second vertebra with a fan-shaped neural process vs. rectangular), the original diagnostic characters by the describers (Sonnenberg & Busch) being not confirmed by Costa … then Killi-Data will keep (conservatively) Nimbapanchax as a subgenus of Archiaphyosemion, not a distinct genus]. 

[addendum (July 1. 2015), : a new work with molecular data processed through the maximum likelihood method by Andrew Furness {ref. Furness, A.I. 2015. The Evolution of an annual Life Cycle in killifish: Adaptation to ephemeral aquatic environments through embryonic Diapause. Biological Reviews of the Cambridge Philosophical Society (Biol. Rev.)} with taxonomic notes on Nothobranchiidae quotes: «Epiplatys maeseni came out strongly placed within the newly erected genus Nimbapanchax» which is opposite to Sonnenberg's original molecular tree ; hence it is more parsimonious not to consider Nimbapanchax as a valid genus (in Killi-Data, it is conservatively kept as a valid subgenus, only)]. 


It might be opposed that researchers need to give names to their research results, but this is not automatic (and Parenti as a professional ichtyologist and Scheel as an amateur ichthyologist have shown an alternative way). It might be opposed that if the new taxonomic generic name is published by a magazine with scientific anonymous reviewers, then it should be accepted as such, but this is not obvious (and it is well known that any author's view is accepted by reviewers as soon as he gives a published reference for his philosophy, e.g. his concept obedience for a genus or a species). It might be opposed that a vote of the board of reviewers of Killi-Data could be a solution, but in practical terms it appeared inaccessible (truely some reviewers soon expressed that this was not their mission implying they would not like to take a position on this very "political" issue).
To make it simple and straight (without caricature), it may be suggested that the old distinction between lumpers (few generic new names with many species for each) and splitters (many new generic names with few species for each) is reappearing with new costumes.

  • When morphological issues were dominant, there were already lumpers (e.g., Scheel) or splitters (e.g., Ahl)
  • When osteological issues became pregnant, with the rising of the cladistic era, there have been lumpers (e.g., Parenti) or splitters (e.g., Costa)
  • When molecular issues became unavoidable, with new techniques from genes, there were immediately lumpers (e.g., Collier) or splitters (e.g., Sonnenberg)
  • When, as today, computer issues become thought as the global "perfect" view, mixing morphological, osteological, and/or molecular data, there are still lumpers or splitters trying to hide with numerical data what remains a strategy to picture their results with a personal consequence of many names (each for big tree branch only as per lumpers, each for each tree sub-branch as per splitters) or not.

Fortunately, to the contrary to old times, no quarrels with personal attacks are raised today from these distinct tactics of lumping and splitting… but this is no better, because researchers ignore each other, even not quoting some results of their "colleagues" provided that the review committee of their concerned manuscript gives the green light for it. Quarrels meant immediate discussion but endless and sterile disputes. No quarrel today and objective neglect mean no disputes, but delayed independant discussions and delayed improvements of understanding. And today for an expert, the analysis of the bibliographic works that are not quoted intently in a given publication or the analysis of the detailed data of a used matrix with no discussion on data previously published by others both mean more on the intentions of an author than in the past with quarrels !

As a consequence, from now on in Killi-Data online, such a splitting description of a genus with a weak morphological-osteological diagnosis, will not be followed until a solid diagnosis is forwarded (even if molecular data are distinctive) and such a new genus will be downgraded to the subgenus level. And this decision must not be viewed as a negative or permanent judgement of the concerned publication or of the concerned researcher(s)… but as a strategy of stability, consensus, and still a spur for in-depth research. This decision is major because in the past any new taxon published was followed without benchmark in Killi-Data (it still remains the case for taxa at the species level, but in the data base there is a note for each taxon mentioning if the move have been approved (confirmed) by other authors in publications, or not).


Obviously, all efforts will be made so that the decisions remains neutral and objective and humble and so that the flow of research works is not impaired. As an example, this same year was published a new generic name by Costa, Notholebias, a split from Leptolebias and it is readily accepted because the diagnoses are well supported and clear-cut.




In total a very important and sensitive newsletter !
Hopefully a boost to our community and a spur to speed up knowledge progress on our (beloved) fishes !


Take care and enjoy the scientific or aquaristic complexity of killifish !

Do not hesitate to ask questions for future Newsletters.

Visit frequently the website!

Thank you for your support over the years.

With my kindest regards.


I am interested in reading other Newsletters, click INFOWEB.