For the purpose of this talk, I am planning to present new data and results on the phylogenetic (evolutionary) relationships of the family Rivulidae based on molecular (DNA) data. The talk will begin by introduction to the phylosophical basis of cladistic systematics. The rest of the talk will be structured in a hierarchical manner, such that the most inclusive group (suborder Aplocheiloidei) will be presented first, with the least inclusive group (the genus Austrofundulus) presented last.
Cladistic systematics looks to elucidate phylogenetic relationships of organisms based on the possesion of shared derived characters, and not on the basis of overall similarity. To briefly demonstrate this principle, one only needs to look at for example sharks, whales and horses. Although there are many over all resemblances between a whale and a shark, such as the fact that both organisms live in water, have fusiform bodies, have fins or fin like appendages, etc., whales are in fact share a more recent common ancestor with (much more closely related) horses than they do with sharks. This statement is based on the fact that, for example, the front legs of horses contain the same bones as do the flippers of whales. Whale flippers have only been modified by natural selection to look like fish fins once the ancestors of whales became fully aquatic. Other sources of evidence are warm-bloodedness in whales and horses, the same cardiovascular system etc. As you will see, this directly relates to killifish, in that traditionally non-annual killies that tend to be fusiform and relatively non-descript have been placed in the genus Rivulus, although in fact members of the genus Rivulus are not closely related to each other, i.e. most species are not Rivulus at all.
Beginning with the main portion of the talk, I will discuss previously published morphological hypotheses of aplocheiloid relationships as put forth by Parenti and Costa.
In the main portion of the talk, I am planning to discuss the phylogenetic relationships within the suborder Aplocheiloidei (order Cyprinodontiformes) as inferred from mitochondrial DNA (mtDNA). The aplocheiloid phylogeny reveals four independent groups of annual fishes, and thus four independent origins of developmental diapause within the suborder Aplocheiloidei. This result conflicts both Parenti’s and Costa’s results. The here presented aplocheiloid phylogeny, however, is supported by geological data. Based on the undisputed theory of plate tectonics, present day land masses have not always had the same positional relationships to one another. Approximately 200 million years ago (MYA) all the continents were joined in one super-continent called Pangea. At that time, Pangea started to break up, first splitting into Laurasia (northern portion) and Gondwanaland (southern portion). Gondwanaland consisted of present day South America, Africa, Madagascar, India, Antarctica and Australia. It is on the first four land masses that one finds aplocheiloid killifishes. In time Gondwanaland started to break up as well, such that ~145 MYA the India/Madagascar land mass broke away from the South America/Africa landmass, then ~100 MYA South America separated from Africa, and some times later India separated from Madagascar. All of these geologic events plus others are reflect in the patters on phylogenetic relationships of aplocheiloid killifishes. This has the implication that although the Indian genus Aplocheilus looks similar to the predominatly west African genus Epiplatys, the two are not very closely related. Rather Epiplatys is closely related to other west African aplocheiloids, the former "Roloffia" species.
In the next portion of the talk, I am planning to discuss the phylogenetic relationships within the family Rivulidae (order Cyprinodontiformes) as inferred from mitochondrial DNA (mtDNA). The rivulid phylogeny reveals two independent groups of annual fishes, and thus two independent origins of developmental diapause within the family Rivulidae. Phylogeny of the family Rivulidae shows high concordance with predictions derived from the geological history of South and Central America. Basal lineages in the rivulid phylogeny are distributed primarily on geologically old areas, whereas more nested lineages occur in geologically younger areas. However, there was little concordance between the molecular phylogeny and currently available morphological hypotheses and existing taxonomies. Based on the mtDNA phylogeny, the genera Pterolebias, Rivulus, Pituna and Plesiolebias are not monophyletic (homogeneous), and do not form natural groups.
In the next portion of the talk, I will be addressing in greater detail the relationships among Austrofundulus populations. As currently defined, the genus Austrofundus contains two species, A. limnaeus and A. transilis. Austrofundulus limnaeus is known from the Carribean coast of Colombia, the lago Maracaibo basin and the Sanare/Tucacas area, both in Venezuela, and the Rupununi savannah in Guyana. Austrofundulus transilis is known from the r?o Orinoco Llanos and the r?o Unare basin. When one takes a close look at the distributio of the Austrofundulus populations, one immediately sees that the populations are isolated by mountain ranges. Based on geologic evidence, one knows the sequence and the approximate timing of the mountain building. If the present day distribution of the Austrofundulus populations is indeed the result of these events, the phylogenetic relationships of the Austrofundulus populations should closely miror the geological events of the region. In the first set of orogenic events ~10 MYA the coastal lowlands of Colombia and the now r?o Magdalena basin from the rest of northern South America. Subsequent to that, ~5 to 3 MYA the Maracaibo basin became isolated from the rest of northern South America. Further mountain building resulted in the formation of the Cordillera de la Costa and El Tigre highlands, respectively, effectively isolating the Tucacas lowlands and the r?o Unare basin from the r?o Orinoco Llanos. The events isolating the Guyanese populations from rest of the Austrofundulus populations are more complex, but past historical events also predict the relationship of the Guyanese populations to the other Austrofundulus populations.
In general each of these geological events cut off the initially westward flowing r?o Orinoco, resulting in a sequential shift of the delta in an easterly direction. The proto r?o Orinoco initially exited through the now r?o Magdalena delta, then throught the lago Maracaibo basin, then through the Chivacoa gap and the Tucacas lowlands, and then through the r?o Unare basin before finally exiting through its present-day delta. These geological events thus sequential isolated the coastal Colombia south of Santa Marta massif, the lago Maracaibo basin, the Tucacas lowland areas, and the r?o Unare basin from the current r?o Orinoco drainage, predicting a pattern of relationships among these drainages. This pattern is clearly reflected in the relationships of the Austrofundulus populations.
In conclusion the taxonomy of aplocheiloid killifishes is complicated and complex. Often killifish taxonomy does not reflect evolutionary relationships, which complicates matters for researchers and can complicate matters for the hobbyist. It will probably still be a number of years and a result of collaboration between a number of researchers before all issues will be resolved. From the practical point of the aquarist, for now relatively few things have changed. However, in many cases such as illustrated on the example of Austrofundulus limnaeus, it is important to keep populations separated and distinct, since in reality we are dealing with different species. In other cases, populations of species are just populations, but in the majority of cases nobody has studied this question, thus we really do not know. When necessary and appropriate, name changes and descriptions of new species will eventually come.