PlanetCatfish.com

....to infer an "accurate" biogeographical history of the siluriformes?

The trees proposed by de Pinna (1998), Diogo (2004) and Sullivan et al (2006) are all pretty different, and I am having difficulty in reconciling the biogeographic patterns.

The Loricarioidei are generally regarded as monophyletic, and as being one of (if not the) most plesiomorphic clades within the siluriformes.

Yet, they are only found in the neotropics. I would expect this most basal clade to have an African and/or even Asian distribution given its potential age.

Can we assume an ancestral distribution and subsequent extinction of the Loricarioidei in Africa, or some sort of event preventing their migration to the "old world"?

If we follow Sullivan (2006) for now, how are the two South American only clades (cet, asp, auc, dor) (pse, pim, hep) nested within Asian and African clades?

For this, would we have to assume the South American distribution and subsequent extinction of ancestral clariids, bagrids, sisorids and silurids etc?

Is it generally assumed that centres of origin can be indicated by the presence of apomorphic or pleisomorphic groups? Authors seem not to agree on this.

Also, has it been resolved to any degree of confidence that the gymnotiformes and siluriformes are together sister to the characiformes, or that the characiformes and gymnotiformes are together sister to the siluriformes? Saitoh et al (2003) summarise the research, but I can't work out which hypothesis they are in favour of.


Anyway, thanks very much to anyone who can be bothered to answer my questions.

I'm no expert but I do enjoy trying to imagine the way evolution operates.
Perhaps the ancestral Loricarid evolved from a Doradid
form?
I know that fish evolution, despite it's diversity in tropical SA, has been remarkably stable compared to many temperate zone fish that have had climatic change functioning as drivers of their evolution.
That stability is well represented in the SA fossil record.

There aren't any Old World catfish that appear to me have much of a chance of playing a role in Loricarid evolution.

Seen this?

IMO and empirically... If you want 'robust', as in 'group likely to be recovered by other studies using different data and methods of analysis', from a molecular phylogeny, then look for a bootstrap tree and collapse all nodes with less than 70% support.

Unfortunately, such a tree will be largely collapsed and therefore uniformative, and what resolution there is completely uncontroversial and already accepted based on morphological evidence - which is why hardly anyone produces such trees.

Oh, and while there's reason to be wary of single most parsimonious trees and single most likely trees, be especially suspicious of trees produced by neighbor-joining analysis - the only reason to use neighbor-joining is because it forces resolution.

Perhaps the ancestral Loricarid evolved from a Doradid form?

The doradids do appear quite primitive with their generally unspecialised body plan, bony cranial shield and plates. However this is not supported by the most studies, which tend to see them as more recent derived forms. Seeing the trees below might help to picture it.

Also, when I say Loricarioidei, this a sub-order, which includes the Loricariidae, Callichthyidae and Trichomycteridae among others. See tree one below.

Seen this?

I have indeed, but I have not noticed any reference to why the Loricarioidei are not represented in Africa/Asia. Am I not understanding something fundamental in phylogenetic biogeography?, or did either they not physically make it over, or they did get over there but no fossil record has yet been discovered.


Sullivan 2006


Sullivan 2006


Diogo 2004

Loricarioidei are not represented in Africa/Asia

Racoll,
They are clearly one of the last new comers in the New World. Note the paucity of Loricarioidei in the oldest (pre-Andes) basins in South America. Corydoras, for example, is completely absence from the huge Magdalena system. There is also only a single (probably representing one of, if not "the" ancestral type) doradid in the Magdalena; Centrochir crocodili. The loricariids present in the Magdalena and Maracaibo basins are all "basic" types (i.e. Panaque, Hypostomus, and a few loricariinae with Neoplecostominae, Hypostominae, and other groups altogether absent).

Clearly the Loricarioidei only experienced a massive expansion of families, genera, and spp after the Andes rose and created the Amazon and Orinoco. Corydoras likely did not exist before this happened as the genus is absent from the old basins.

So, I think the answer is that they are absent from Asia/Africa either because their ancestral type did not exist that far back, or was unsuccessful in adapting. It would appear that it was only the collision of the two tectonic plates in the Pacific that formed the Andes that allowed this group to come into their own.
-Shane

PS My question is what the heck happened to the cyprinids in South America? They "rule" North America, Africa and Asia but missed out on the largest ,most spp diverse basins in the world. Go figure.

PS My question is what the heck happened to the cyprinids in South America? They "rule" North America, Africa and Asia but missed out on the largest ,most spp diverse basins in the world. Go figure.

Thats easy...... the corydoras ancestral type ate them all before evolving into the nice inocent looking fellas that we see today so that know one would suspect them, it's obvious really.

But seriously though a coincedence that you ask this question Racoll, I am currently reading 'The Ancestor's tale' by Richard Dawkins and was going to post pretty much the same question only with the addition of how did the corydoras get to trinidad.
I hadn't thought about the cyprinids though, but it is glaring now that it has been pointed out to me.
I too am no expert but I feel that an extinction senario sounds apropriate to me (regarding Loricaridae) and would say that a lack of fossil discoveries doesn't mean that they are not there and may be down to the types of habitat I would have thought a loricarid ancestor would live in may not have had very good conditions for fossilisation to occur, not much sediment deposited in a rushing river I wouldn't have thought.
I cannot see the trees that you have posted as I am on a clockwork computer but I imagine that the loricarids that do now live in slower habitats with good conditions for future fossilisation would have evolved from a rheophilic species rather than the other way around.
Please stop me if I'm talking Bollox.
Matt

PS My question is what the heck happened to the cyprinids in South America? They "rule" North America, Africa and Asia but missed out on the largest ,most spp diverse basins in the world. Go figure.

John Briggs has a nice theory about that. Check Journal of Biogeography (2005), volume 32, pages 287-294 for more.

John Briggs has a nice theory about that. Check Journal of Biogeography (2005), volume 32, pages 287-294 for more.

There is a lot I don't really get about Brigg's theories. He seems to be out of sync with other authors. I don't follow how he thinks that presence of apomorphic (derived) groups indicates a centre of ancestral origin.

I prefer the explanation presented by Saitoh et al (2003) where it is proposed that the Otophysan ancestor originated in the Laurasian part of Pangaea at about 250ma (in the early Triassic) and invaded Gondawana. When Gondwana broke from Laurasia the Gondwanan group evolved into the siluriforms, characiforms and gymnotiforms, and the group left in Laurasia became the modern cypriniforms. The gymnotiforms never made it to/survived in Africa, but the siluriforms successfully radiated into Africa and then into Asia when it rejoined.

However, putative landbridges between Laurasia and Gondwana in the early Cretaceous probably suggest the siluriforms invaded Laurasia from Gondwana much earlier (this could also account for the Eocene Hypsidoridae found in the USA). However for this to work, we must accept that the Loricariodei and Diplomistidae didn't invade/went extinct in Laurasia, and that the cypriniforms did the same with regard to South America.

Diogo (2004) presents a Pangaean origin of catfish, but this is partly based on the sister grouping of the Aspredinidae with the Erethistidae, a relationship not found by Hardman (2005) or Sullivan (2006).

was going to post pretty much the same question only with the addition of how did the corydoras get to trinidad.

Compared to the above, this is dead easy. Apparently Trindad separated from South America around 11,000 years ago, and Corydoras were essentially modern at 50ma.

I imagine that the loricarids that do now live in slower habitats with good conditions for future fossilisation would have evolved from a rheophilic species rather than the other way around.

This is probably correct as all the extant basal groups of the Loricariidae are steep gradient stream species. However there are many instances of evolutionary reversals and range shifts, which is something to bear in mind.

Clearly the Loricarioidei only experienced a massive expansion of families, genera, and spp after the Andes rose and created the Amazon and Orinoco. Corydoras likely did not exist before this happened as the genus is absent from the old basins.

The Andes were supposed to have risen from about 90ma(?), and the earliest record of Corydoras was from 50ma, so it is possible, but as Matt says, absence of evidence is not evidence of absence.

The Andes were supposed to have risen from about 90ma(?)

Seems I got this completely wrong. The Eastern Cordilleras are reckoned to have cut off the Magdalena at about 12ma (Albert et al 2006). This means the Corydoras existed well before this event.

They state:

"In terms of species composition, the ichthyofaunas of all river basins in NSA were largely modern by the time of the Late Miocene tectonic events that dissected the landscape."

Hmm..I've changed my mind completely with this now that I am off the clockwork computer and can see the doodles.
It seems that the question was not read very well by myself and therefore was talking absolute bollox.
I was just wondering, the fish that are covered by the 'Loricarioidei' label in modern examples anyway, seem to me to be fish that 'stick with what they know' as in, if they are designed for fast water that is where they stay if they are designed more for slow then that is where they stay, e.g astroblepidae- fast cool water
Ah but, there are loricariidae that live in all conditions fast, slow, hot or cool you say and corys' for every occasion!
That is not what I mean, if we look at the L046 we find a fish that lives in well oxygenated warm water i.e the Rio Xingu
However if we look at Phractocephalus hemioliopterus we find a fish that actually goes anywhere and I mean an individual fish.
I think that you would find this with the other ones that made it 'back to' South America, Doradidae Aucheniperidae.
The Larger Doradidae migrate to breed and so do Big Pims.
My point is after all that is that these fish are 'adventurous' in their nature, they go places on a local scale and their ancestors went places on a global scale.
The Loricarioidae are the couch potatoes, they liked it there and they are going to stay there, mumble mumble...new fangled where are you going? It'll all end in tears you know! and off the ancestor of the others went, literally, around the world, leaving the Loricarioidae to nestle into their little niches returning from their worldly travels much changed to find their Loricarioidae mates still siting in the same chairs down the pub whining on about the same crap as when they left them, if you get my meaning.
They never went to Africa because they could see no need (in an evolutionary sense) they had everything they needed at home.
What do you think?
Matt

grokefish wrote:I was just wondering, the fish that are covered by the 'Loricarioidei' label in modern examples anyway, seem to me to be fish that 'stick with what they know' as in, if they are designed for fast water that is where they stay if they are designed more for slow then that is where they stay, e.g astroblepidae- fast cool water
Ah but, there are loricariidae that live in all conditions fast, slow, hot or cool you say and corys' for every occasion!
That is not what I mean, if we look at the L046 we find a fish that lives in well oxygenated warm water i.e the Rio Xingu
However if we look at Phractocephalus hemioliopterus we find a fish that actually goes anywhere and I mean an individual fish.
I think that you would find this with the other ones that made it 'back to' South America, Doradidae Aucheniperidae.
The Larger Doradidae migrate to breed and so do Big Pims.
My point is after all that is that these fish are 'adventurous' in their nature, they go places on a local scale and their ancestors went places on a global scale.
The Loricarioidae are the couch potatoes, they liked it there and they are going to stay there, mumble mumble...new fangled where are you going? It'll all end in tears you know! and off the ancestor of the others went, literally, around the world, leaving the Loricarioidae to nestle into their little niches returning from their worldly travels much changed to find their Loricarioidae mates still siting in the same chairs down the pub whining on about the same crap as when they left them, if you get my meaning.
They never went to Africa because they could see no need (in an evolutionary sense) they had everything they needed at home.
What do you think?
Matt

I've been out of practice reading any taxonomic papers or reviews since completing my PhD on gobies 15 years ago, but this discussion is close to my heart as I've always been fascinated by catfishes; however I'll state in advance that I've NOT read the papers quoted so my opinion is probably useless- apologies to the real scientists out there. Matt's discussion above seems to have at least a ring of sense to it- all the South American groups that are separated from the basal clades contain species which have a tendency to range widely as individuals, often containing large species which tolerate estuarine water. Thus a radiation from an ancestral catfish originating in South America seems consistent, with other ancestral groups being somewhat more sedentary.

The Ariids are easy enough to explain as they contain lots of marine species; doradids and pimelodids also have many large species capable of swimming large distances during their lifespan and tolerating slightly brackish water. Aspredinids wouldn't fit my idea of an active colonising group, but these also have euryhaline tolerances- something that is lacking in many of the basal South American clades on Diogo (2004). So maybe a tolerance for brackish water was a prerequisite for catfish radiation?

Maybe I'm missing something but Sullivan (2006) doesn't seem to make any observations about sister groups at a "bottom level"- there seem to be 11 unresolved groups on this scheme.

A quick and completely unrelated question... does anyone know what software Sullivan used to produce those pretty trees?

I'm in the process of prettifying a big tree for publication, and it looks like whatever he used might be just the ticket.

MacClade 4.0 I think.

racoll wrote:

was going to post pretty much the same question only with the addition of how did the corydoras get to trinidad.

Compared to the above, this is dead easy. Apparently Trindad separated from South America around 11,000 years ago, and Corydoras were essentially modern at 50ma.

There's actually some good evidence that suggests that Trinidad may have been connected as recently as 3-4,000 years ago, with the Gulf of Paria as a huge wetland or landlocked sea. In addition, there are fish species that are restricted to the southwestern peninsula of Trinidad, which suggests that introductions are still happening. The Gulf of Paria is only around 23 ppt, and when the Orinoco is in flood the Columbus Channel (which is only about 10 miles wide) is probably a lot lower. If capybaras and alligators can make it across on "vegetation islands", I wouldn't be surprised if fish can.