As early as 1828 Wagner recognized a suprageneric category for the group known to us today as the Cyprinodontif-ormes. In 1835 Agassiz erected the Cyprinodontes as the family containing the genera Cyprinodon, Lebias, Molinesia (correctly spelled Mollienesia), Poecilia, and Anableps. By 1883 the term Cyprinodontidae was in general use for 30 genera and 130 species, as reflected in the work ofJordan and Gilbert. In his classic work The Cyprinodonts, published in 1895, Gar-man arranged the many genera of the Cyprinodontes into eight subfamilies, and while he erroneously included the characin genus Neolebias and the cyprinid genus Fundulichthys, his systematic view of the constituents of the Cyprinodontif-ormes is about the same today. This is reflected in the work of Berg, who formally erected the group as the order Cyprin-odontiformes in 1940.
In 1964 Rosen placed the Cyprinodontiformes in the su-perfamily Cyprinodontoidea in the order Atheriniformes. As of 2002 the Cyprinodontiformes were considered a natural group (i.e., a monophyletic group) most closely related to the Beloniformes, the order containing halfbeaks, medakas, needlefishes, sauries, and flyingfishes. Rosen took Garman's eight subfamilies and ordered them into five families, the Cyprinodontidae for all the oviparous genera, the Anablepi-dae for the viviparous genus Anableps, the Jenynsiidae for the viviparous genus Jenynsia, the Goodeidae for the genera of the viviparous splitfins, and the Poeciliidae for the viviparous genera with gonopodium. In 1924 Hubbs had argued for the placement of Anableps and Jenynsia into the family Anablepi-dae, a view not adopted by Rosen. The five families of Rosen correspond today to the order Cyprinodontiformes. In Rosen's work there was a clearly defined family-level separation of the lineages into viviparous and oviparous. Essentially, the five families were thought to be related, but there were no proposals about the details of the relationships among the five families or their genera.
It was not until the iconoclastic work of Parenti in 1981 that relationships for the cyprinodontiform genera and their families were proposed. Shared derived characters (evolutionary novelties), not primitive characters, were used to define the various evolutionary lineages in a systematic phylogenetic analysis of the order. Almost all the known genera were reevaluated, and a comprehensive cladogram was constructed to illustrate the proposed interrelationships of the genera and the families into which they were placed. For the first time the order Cyprinodontiformes was defined using only derived characters, consisting of a suite of various oste-ological features and, in general, a long developmental period and early breeding habits. In the new ordering the viviparous families did not form a monophyletic group, because some live-bearers turned out to be related more closely to oviparous species than to other live-bearers. There were numerous tax-onomic and nomenclature changes as the result of the restructuring of the genera as well as the proposed scheme of their relationships.
Three taxonomic arrangements were proposed, wherein viviparous genera were deemed to be sister taxa to oviparous genera. This kind of relationship had not been contemplated previously and constituted a paradigm shift within cyprin-odontiform systematics. The viviparous genera Anableps and Jenynsia were recognized as sister taxa because of shared characters in their reproductive biology. These, in turn, were considered the sister group of the oviparous Oxyzygonectes, which had been aligned with Fundulus. These genera constitute the Anablepidae. The viviparous family Goodeidae, with 17 genera, commonly known as splitfins, was realigned with the sister group formed by the oviparous genera Crenichthys and Empetrichthys. This group now constitutes the family Good-eidae, which may be considered to have two subfamilies, the Goodeinae and the Empetrichthyinae. The Poeciliidae, with 27 genera, was placed in a clade with the oviparous African lampeyes, Aplocheilichthys, and its related genera and the
oviparous genus Fluviphylax. That group, in turn, was hypothesized to be the sister group of the Anablepidae.
As a result of these revisions, the killifishes, that is, the oviparous cyprinodontiform genera, are no longer considered to constitute a monophyletic group in the scientific sense, but in the vernacular the term is used commonly. It also should be noted that the Cyprinodontiformes were subdivided into two suborders, the Aplocheiloidei and Cyprinodontoidei, both of which spanned the continents of South America and Africa, with interesting zoogeographical implications.
Four major revisions of the Cyprinodontiformes have been published since 1981. Using both molecular and osteological methods, overall cyprinodontiform taxonomy was revisited, the phylogenetic relationships of the Old World and New World aplocheiloids were revised sharply, and a major restructuring of the Poeciliidae was undertaken. An additional molecular study of the phylogeny of the family Rivulidae and its two subfamilies, the Rivulinae and Cynolebiatinae, was published in 1999, and this may lead to a radical taxonomic and nomenclatural revision of these two subfamilies.
The fossil record does little to illuminate cyprinodontiform origins. The earliest fossils, the cyprinodontid Pachylebias and Prolebias, both found in Europe, date to the Oligocene epoch, 25-40 million years ago (mya). Most fossils date to the Miocene, 10-25 mya. The fossil record is relatively recent compared with the inferred history of the Cyprinodonti-formes. The phylogenetic relationships of recent killifishes and live-bearers accord well with the realities of plate tectonics and the breaking up of Gondwana, the single supercontinent formed by present-day South America, Africa, Antarctica, Australia, New Zealand, Madagascar, and India plus an assortment of other small plates not part of the continent of Laurasia, which was positioned to the north of Gondwana.
The distribution of the Cyprinodontiformes on today's widely separated pieces of Gondwana argues for a very ancient origin of this order of fishes, more than three times the age of the oldest-known fossil killifishes. The killifishes and live-bearers originated on Gondwana and were contempora neous with the dinosaurs. Plate tectonics (continental drift) carried these freshwater fishes to their present locations on a journey that began with the breakup of Gondwana 140 mya. The mountain killifishes, genus Orestias, found in the Altiplano regions of Peru, Bolivia, and Chile, were in place long before the rising of the Andes lifted them to their lofty positions. In general, one can expect some dispersal within zoo-geographical areas, but the distribution of freshwater fish families is a result mainly of continental drift.
Before plate tectonic theory, freshwater fish distributions presented great puzzles, the answers to which were sometimes fanciful speculations. For instance, both the African killifish genus Aphyosemion and the South American killifish genus Rivulus were placed in the same subfamily, the Rivulinae. If one thinks of the continents as always being in the same positions, how can the freshwater fishes of these two continents possibly be related, since there is no way small freshwater fishes can swim from Africa to South America? One of the more fanciful hypotheses, ludicrous by today's more sophisticated standards, was that a series of islands spanned the Atlantic Ocean from South America to Africa. The freshwater fishes supposedly swam from island to island, thus accounting for the separation of their families. Then the islands conveniently disappeared without a trace. Another theory held that the continents had moved apart because the earth was expanding. (Two marks on a balloon grow farther apart as the balloon is inflated.) That theory suffers from a lack of any plausible mechanism, although in its early days the theory of continental drift was subjected to the same criticism, a criticism that was answered by the now widely accepted hypotheses of seafloor spreading and subduction.
The systematics of the cyprinodontoid family Poeciliidae parallels the biogeographical situation outlined for the aplocheiloid genera Rivulus and Aphyosemion, since part of the Poeciliidae is South American and part is African. There, too, plate tectonics offers a satisfying explanation of the biogeog-raphy of that group.
The taxonomic evaluation of the Cyprinodontiformes is far from complete, and one may reasonably expect many more far-reaching revisions. Many new cyprinodontiform species are being described and will enrich our understanding of this order. In 2000 Lazara pointed out that the number of species described since the first killifish was cited in 1766 by Linnaeus has increased exponentially.
Was this article helpful?