Killi
ENCYCLOQUARIA . SCIENCE
The Art and Science of Aquaria





Before the Continents Split
Where Killifish Came From


Do you ever wonder why there are no killies in the far Australia which has Rainbow fish instead? Or Japan that has none but has the almost-killi-like Ricefises? Killifish originated in an area that was then where the dividing lie between what would be Africa and North America stands and they radiated outward from there. They never got as far as Australia or if they did were unable to last in the same way Aponogeton flowers show up in the fossil record in Greenland, Newfoundland and Peru but they were unable to survive there.



Killifish are between Rainbowfish and needlefish in the tree of (Piscine) life, the two share a common ancestor 75 million years ago, which also share a common ancestor with the Engineer Gobies and there were no further descendants of that clade. Thus the rise of the Cyprinodontiforme fish happened right after the rise of the Silversides, that we see as Rainbowfish, and was followed by the rise of the Needlefish and then Engineer gobies. At the 100 million year mark all these fish come from a common ancestor. Therefore the ancestor(s) of all killifish have a fairly fixed point in time.



We do not know with any great precision when killifish and rainbow fish swam from the reefs and estuaries into freshwater and while there are remnant killifish and rainbow fish species at still inhabit salt water (the the most notorious of which is in Nevada, US, the Devils Hole Pupfish that has 91F hyper-saline water) most occur in freshwater with some found in pure rainwater, a very delicate chemistry.

But we do know that they did and in broad strokes, something like Pachypanchax playfairi, a very long time ago, swam into a river from the sea, whether to escape predators or to find food does not matter nor does it matter if it made it 100 years and then died, because another one will try it and over time enough try that a few succeed and establish breeding colonies inland, and this as the start of the rise of the Cyprinodontiformes fishes all over the world a the continents split apart around 100 million years ago, shown here.



Animation of the break-up of the supercontinent Pangaea and the subsequent drift of its constituents, from the Early Triassic to recent (250 Ma to 0). In contrast to the present Earth and its distribution of continental mass, much of Pangaea was in the southern hemisphere and surrounded by a superocean, Panthalassa. Pangaea was the most recent supercontinent to have existed and the first to be reconstructed by geologists; here is more detailed explanatoin of supercontinent formation from Steven Dutch, at the University of Wisconsin.

At the time this continental breakup began, something like this swam for whatever reason from saltwater into a freshwater river, the same sort of shallow river that killifish are found in world over, these are river not lake fishes and they are overwhelming most often found in very shallow streams often no more than an inch or two deep.



Rainbowfish on the other hand and not denizens of shallow water and their upturned pectoral fins suggest a life spent closer to the surface of the water they live in than the bottom. We do not see this evolutionary adaptation in anything but surface dwelling fishes and this implies a certain depth.


Glossolepis multisquamata


So, killifish lost the front dorsal and became less flat and tall and more round as they adapted to riverine life instead of a pelagic marine existence. This took about 20 million years.

So if you wonder why sometimes Aphyosemion herzogi and Aphyosemion riggenbachi have tails suggestive of gularis youneu needn't wonder: they inherited them from an ancestor they and rainbowfish have in common. The genes were always there, other gees that expressed them had not yet been turned on, that's all. This also explains why sometimes South American and African fish look nearly identical as the case of the African Aphyosemion striatum and the South American Rivulus mahadensis.


It's also the reason Gularis are not the only ones with those tails.

In the big scheme of things something like Pachypanchax playfairi, pelagic marine fish swam upstream into a freshwater stream between 100 and 250 MYA and over time turned into something like A batesii. From these several radiations outward in at least 4 major epochs resulted in all derivative forms of Aphy,Fpx and roloffia (Callopanchax is to Scriptaphyosemion as Raddaella is to Aphyosemion).

How exactly the Epiplatys, nothos and lampeyes arose is another question but as you can see the plasticity exists for nothos to become lampeyes,here's one doing it right now.

Now, whatever P. Playfairi-like fish did swim upstream would also have ended up in any big lake that form so one wonders if the one that ended up in a lake if perhaps that one would have changed the least. Keeping in mind the rainbowfish split off just before the killies do in the tree of life then how distant can L. tanganicus be from the P. playfairi like fish all killies are descended from?

Playfairi is more rounded and suited to more shallow streams an adaptation they share with nearly all other killies except te pelagic lampeyes. This killies become as lampeyes in rovers and small lakes and like Aphyosemion and Nothobranchius where water is scarce and even seasonal or ephemeral.


Lamprichthys tanganicanus: a killifish

Rhadinocentrus ornatus: a Rainbowfish


Two things have always stuck in my mind: that Scheel referred to Epiplatys as Aplocheilus in his second book and that rainbowfish eggs are like killi eggs.

Now, Aplocheilus and Epiplatys are still distinct,that idea did not resonate with many.But,the fact you had to think about it a bit to show why it can't be true and that somebody of Scheel's caliber would think so does give cause for alarm.

Here's another. Were told all killifish are descendants of a pelagic marine species that swam inshore around 100MYA when the continents split apart and that it probably resembled Pachypanchax more than anything else and not hard to look at Raddaella batesi and note it's not that different really and bears similarity to some rivulus on another continent, and more than a few actually.

But the cylindrical body shape of Pachypanchax is not what a pelagic marine species would look like and we have to ask ourselves what were the species of fish that killies split off from and the answer is the Rainbowfish.

So, in the same line of thinking that suggests all Epiplatys are really Aplocheilus, at one level up in the taxonomic tree one could ponder the idea rainbowfish and lampeyes are the pelagic form while killifish are the forest form.

That would also explain why killies ar found all over the world expect australia, and why rainbowfish also only occur in Australia or at least behind the Wallace line.

It seems to me the fuzzy line between rainbows and cyprinodontiformes fishes gets pretty thin. The idea that something like Pachypanchax is the ancestor of all killies doesn't resonate well with me as it's not a pelagic fish, it's adapted to shallow rivers and the color pattern may be similar but the body shape of a pelagic ancestor would look more like a lampeye and less cylindrical. It also stands ot reason if any of those ancestral forms swam into a lake and not a river then it might speciate in color but less in body that riverine Cynodonts, thus,can we find a lampeye and ar rainbow that are similar?

Here's the only endemic lake dwelling Cyprinodont,it swam into Lake Tanganyika and stayed there. It's not found anywhere else.

This rainbowfish from australia (Rhadinocentrus ornatus "Carland")... how different really is it from tanganicus?



Is the gene that makes the orange here the same as the "cameronense orange" that shows up in gularis?

If somebody sold that rainbow as a new Lamprichthys species how believable would it be?

How hard is it to believe tanganicus and this rainbow had a common ancestor?

It's a lot less changes,for example, than exiguum has to go through as derived from Raddaella.

So maybe "killifish" or "fish of the creek" isn't accurate. Maybe when the continents split rainbows headed for freshwater and we have pelagic rainbows all over australia and africa and only one left in South America and none in Asia but we have forest rainbowfish all over Africa and South America and in a few places pelagic and forest forms intermingle, that's why we'll have an aphyosemion, an epiplatys and an lampeye so often found together, each state out their part of the niche: epiplatys at the top, Aphyosemion at the bottom the pelagic ones drying in the bright sunshine. Still all the same fish though, just adaptations to their particular niche.

So I'm suggesting rainbows and killies are closer to each other than killies are the the fish that evolved (as a clade) after killies, the needlefish. Two foot long pencil thin gars are not going to be mistaken for killies anytime soon but I do find it somewhere between alarming and disconcerting that a killie and a rainbow could be this similar yet so far apart and convergent evolution only goes so far here.

So one thing that came out of Parenti (1981)* was a shift in thinking from "all killies descended from Aplocheilus" to "all killies descended from something very like Pachypanchax."

So you can go for a couple of decades looking at Pachypanchax all weird. How can they be the key to anything?

Then you find this monster, swimming around creeks of Madagascar, the land the time forgot and suddenly it's far less of a mystery.

Look at it carefully, what does it look like? Blue fish, red spots, yellow fins. Sound familiar?


These red fish are also found in same stream.

So, as we ponder the diversity of color patterns killifish have evolved into perhaps we should give more consideration to the color patterns they evolved *from*. It's a lot easier to explain mimbom and schmitti knowing these things may have swum inshore in two different places in the tethys sea that would later be pushed together to for the continent of Africa. The top one, west of the Dahomey gap while the one that came up fro the "south" was all points east of the Dahomey Gap.

I think the idea any killies evolved from any other is specious, you can make any one out of any Pachypanchax easily, and in fact it may be the only way to explain killi genes. We'll see as we learn more about what they are.

The other great question is where did the ancestors of killies come from and where do they come from. Short answer: the grey mullet. Ish.

You can read Parenti 1981 online at the Smithsonian now, Parenti is curator of fishes there presently.

Parent 1981 where, at least in my mind, biogeography really came of age in the study of killies much in the same way chromosome and genetic analysis came about in Scheel's time. ROTOW in 1970 barely mentioned biogeography whereas in 1981 Parenti wrote at length about it.

And that's good, it does help explain a great deal.

However one thing I do not see anything written about and perhaps I just haven't looked hard enough and that's about the soil. We hear mention of leafy basement liter or a rocky or muddy or peat filled pond or stream but that's really al.

Consider killifish have been where they are for about 100 million years plus or minus 50my. In that time many changes in genes and physical appearance occurred, some to a greater extent than others and some at a faster rate than others.

And we tend to look at biogeography as the overarching physical force a work - the environment, we call it, but there are also forces at plan on the genes.

In the barbs there is a genus that has twice the chromosomes of a sister genus, and another that has four times that. In Cryptocoryne one species (COR) has a different number of sets of chromosomes whether you find it i Makaya, Thailand or Borneo. So, the number of sets of chromosomes can double.

But changes can also occur to genes from random means and radiation is the working definition of random, it's the only true source of randomness in nature. You can't generate random numbers, only pseudorandom numbers. It's an issue in mathematics.

The nature of these random mutations are mostly various forms of high energy radiation and cosmic rays are the ones that do the most damage to the DNA molecule. The amount of cosmic rays hitting the earth is limited by the earths magnetic field and that broke up about the time the last great ice age started and that's well in the rise of killies timeframe. OS there's that, one of the reasons for the increased rate of speciation in killies, cichlids and rainbows at the very least is the increase rate of random mutation from more and higher energy cosmic rays hitting earth.

Second, and this is what I find odd is I do not ever see it mentioned but the soil chemistry of the coastal rain forest varies greatly and differing percentages of minerals in different countries ay affect genes or *expression* of genes if a specific mineral is needed to form a catalytic precursor. A good example of this is selenium, found in deposits half a mile thick in Senegal as selenium phosphate, also found in Bolivia and Gabon.

It may turn out varying percentages of exotic trace minerals may alter the efficiency of certain chemical reactions this up (or down) regulating genes that control color and color pattern. It may turn out if you swapped fish from say Makokou and inland Cameroon they make "change into each other" as a reflection of the soil types. This is may be soil as well as biogeography that affects the genetic and thus phenotypic expression of color and color patterns in at least killifish.


Pollutants can cause a reduction of the chromosome count

This is one possile explanation for the Raddaella species: kunzi, batesti and splendidum is one evolved from the other. In this photos it's possible to imagine BAT over time morphing into KUZ and SPL (morederived forms in Cyprinodontidae tend to be omre spender in Aphysemion) but the reverse is not as easy to imagine.

What if instead of KUZ being derived from BAT (which is not impossible) then instead KUZ, BAT and SPL were derived from a common ancestor not a killifish? That is we assume the differences in these species occurred as a result of differentiation of the species as killifish but what id the actual difference was three different color morphs of the common ancestor that wasn't a killie?

The point this raises is if this is true then what is te explanation for the vat difference in color patterns within Raddaella: it's easy to explain them away by assuming they evolved from each other but what if they were reacting to local environmental conditions viz a viz the soil. The common ancestor to killies and rainbows is Anthias it might be possible for genes that are only expresses in marine biota may be repressed in freshwater and in areas of unusual mineral contents perhaps that resembles the trace element profile of the sea then perhaps these unregulated minerals not normally found in all areas (and Selenium ie one such exampe, it's high in areas of coastal Gabon and Senegal only along the coast) cause the genes that express colors to up-regulate foooling them into as it were, into turning on marine color pattern genes, long since inactive since the Anthias days. This idea is supported by the rather tenuous idea that L. tanganicus may have a higher degree of color compared to a lot of lampeyes owing to the mineral content of the water by this same mechanism.

We can tell how far evlved or as we say "derived" killies are with conventional physical metrics, bones and dna. There's a great variance in Aphyosemion for example with Radaella having an absurdly small number of chromosomes and Diapteron having the most,it's something on the order of 8 to 40. Now, chromosomes can be reduced in number from environmental stress but as things become more derived the tendency is to increase the number of chromosomes and this is always going on. A lot of killie species have two or more populations with different numbers of chromosomes even today - the process never stops.

n=18 and 2n=36 - Scheel's fish India n = 19 and 2n = 38 chromosomes." [1978 Khuda-Bukhsh] Vietnam 2n=40 [1992 Magtoon]

so that's one way. Also as killies adapt, the requirement for evolution their jaws change to reflect feeding habits in their new habitat and this should be comparatively quick as their ancestors were pelagic marine fish but as "fish of the creek" they now live in tiny amounts of water and feeding will be specialized compared to a pelagic existence thus structural changes in the jaw occur. From this we can tell more and less derived forms.

Parenti did a thorough examination of the jaw bones of all killifish genera. Not species, genera. The potential gotcha is if we've defined genera improperly this would not be representative of all genera if not all species were examined. The converse though, new general may rise out of this, and they did: Fundulopanchax was raised from a subgenus of Aphyosemion to a full genus. Aphyosemion and Fundulopanchax had different ancestors.

In a nutshell the slender bodied Aphyosemion and the more stout bodied Fundulopanchax are thought now to have evolved from different lines.

Right. So, ever see a picture of what that common ancestor was? The idea it may be some long extinct fossil relic is an obvious conclusion to which I saw: coelacanth.

Parenti states flat out you can't have any assumptions going on and the idea Aplocheilus was the ancestor to all killies was based on metrics. Parenti's work with jawbones of all killies was monumental and gave new insight and showed aplocheilus was a derived form from something pachypanchax-ish.

So, what I'm suggesting here is there were provably at least four species in two clades in the right part of the Tethys sea at the right time such that it's not impossible that killifish could have come from what is now pachypanchax but also from the other fish there at the time. This hypothesis would explain the differences between aphyosemion, a stout fish, blue with red dots and yellow fins and a stout pachypanchax like this still exists today. And in fact a lot, quite a lot of color patterns of west african killies show up in extant Pachpanchax but a lot also show up in the other clade that we now call rainbows but they're having trouble over in rainbow land because the ones in madagascar and one part of AU aren't the same as the rest in the same way it's not obvious to see how lampeyes fit in with the rest of them.

So, what I'm suggesting in response to Parenti saying, roughly "killies come from Pachpanchax" them I'm merely adding: 1) And that thing may have been a collection of things: what if that hypothetical ancestor came in 5 different color morphs. This if we look for the origin of color patterns in killifish there's no real reason to think we'd actually find them, they have have inherited them.

2) One common ancestor is to my mind a flaw in thinking. What it it wasn't one grey fish tat turned into all killies what if it was 6 color forms of the same fish? What if it were two closely related species instead?

Both killies and rainbows have a number of species that just don't fit in anywhere else and there me a whole species grouping between the killas and the rainbows we missed that includes the rainbows that don't fit in, the pseudomugil rainbows, the diopters of rainbowland and the madagascar rainbows. Tanganicus may belong here and that, that opened up a can of worms.

Of course this is highly speculative and not based on any hard evidence and I'm really just thinking out loud. But, at some point Jouke van Der Zee will be finished with his review of lampeyes and we'll know a bit more and have some more pieces to the puzzle.

So the best I can say for this idea is it's self-consistent. I think it's actually easier to imagine some killies have come from Bedotia and not Pachypanchax, there are a number of species of Bedotia and some of them are turning into bivs you'd think by looking at tem plus they're the only source of an trilobate tail and the walkeri orange/fallax tail. I think it's worth looking at. ymmv. Constraints
In trying to analyze evolutionary trends within killifish we mst distinguish between what is possible and what is not possible. First, these are forest fishes for the most part and could not exist until not just forrests but biodverse forests with bugs falling off trees all the time. Another curious coincidence is how the evolution of the human genome in Africa might hpothetically mimic the radiative adaption of killifish throughout Africa the premise being that man, before he was man, was a forest dweller with a need for clean fresh water. Where you find man, you'll also find killifish, we all love the shoreline.