Bugs in Occurrence data on River Tuy, etc.

[Shane]

This has a lot of potential, but will also be a lot of work (in addition to the massive amount already done by our hardworking Mod Mats).
I ran a couple of queries:

Tuy: Of the 10 spp listed, 3 are not found there at all (The Chaetostomafrom the Rio Zuata, Epapterus, and H. plecostomoides). Also, temps are different for most of the spp. They should probably be standardized with the exception of really large drainages. Also, and Jools and I have discussed this, it gives the idea that the listed spp are the only ones present, instead of representing the truth; they represent a small proportion of described spp (this is true for all drainages).

Rio Guarico: Returns two hits, but there are a dozen popular aquarium spp from this river.

-Shane

[MatsP]

This has a lot of potential, but will also be a lot of work (in addition to the massive amount already done by our hardworking Mod Mats).

Thanks, and thanks for reporting the bugs. I'm moving this to the "Suggestions and bugs" section, as that's where it belongs [that way, I can check it and move it to Resolved when it's sorted].

I ran a couple of queries:

Tuy: Of the 10 spp listed, 3 are not found there at all (The Chaetostomafrom the Rio Zuata, Epapterus, and H. plecostomoides).

C. sp Rio Zuata is listed in the Cat-eLog as

Found throughout feeder streams of the Rio Tuy and in feeder streams of Lake Valencia.

Similarly for Epapterus and H. plecostomoides (it explains that the latter hasn't been found for several years - is this the same for the other species you list, i.e. they have once been in that river, but can't be found any longer).

I can only go by the information I've got - which is a combination of existing Cat-eLog (which unless it's obvious to me it's NOT right, I take at highest priority), CLOFFSCA and Fishbase (second priority, despite many errors probably better than "aquarium books"), as well as Catfish Atlas and other books I own. We probably should discuss this in the Bugs section, so I'm splitting this topic and moving it to bugs.

I've never been closer to Venezuela than the Texas coast-line (around Corpus Christi is the furthest south I've been in the Americas) - which is quite a lot too far away to understand what fishes live in which rivers!

Can you enlighten me on the correct distribution for those fishes (and of course, add let me know of any other species in the Cat-eLog that are from this river).

Also, temps are different for most of the spp. They should probably be standardized with the exception of really large drainages.

Also, and Jools and I have discussed this, it gives the idea that the listed spp are the only ones present, instead of representing the truth; they represent a small proportion of described spp (this is true for all drainages).

Both of the above are things I've been looking at previously (I went through and made a list of rivers and temperatures - it varied quite a bit, sometimes so much that the lowest max was below the highest min for a small-ish river).

It would be possible, with this new data, to acutally list the temperature range used for species that live in the same river, and find the "out of range" ones to adjust them. I don't think that's a priority right now, tho'.

The second problem can be solved by adding some extra text to the result page, explaining that "distribution data is point-samples based on current 'best knowledge', and new species not-yet-found are obviously not present in the list, neither is fish that are outside it's presently known distribution"

Rio Guarico: Returns two hits, but there are a dozen popular aquarium spp from this river.

I agree, the guarico was only recently added (body of water ID no 476 of 515 or so), so I have some more fishes to add to this river.

Using my "secret" search for rivers in the "distribution text field", it lists 4 species from this:
Corydoras elegans (mismatch - it's Aguarico, which I'm pretty sure is a different river)

Corydoras habrosus
Hypostomus sp. (L192)
Stegophilus taxistigma

So can you list any other species from this river that are in the Cat-eLog?

And of course, one of the problems here is when a fish is listed as "Rio Orinoco basin", and someone searches on, say, Rio Ventuari - it may not be listed in the distribution as Rio Ventuari, but that's doesn't stop the fish from having a valid distribution in Rio Ventuari - the other way around works if you don't select the "no tributaries" tick-box - it will list all rivers that are connected to Rio Orinoco in the database (which may or may not be correct, of course!)

--
Mats

[Shane]

Mats,
I am not sure that these are "bugs" as much as just a lack of available information.

1) Listing in the Cat-elog for Chaetostoma sp Rio Zuata is incorrect (not sure why). It is restricted to the Rio Zuata. Rio Zuata connects to the Guarico, Guarico to the Apure, Apure to the Orinoco.

2) H. plecostomoides is restricted to the llanos and has never been recorded from the Tuy. H. plecostomus has been recorded from the Tuy (typo?). Check out Lilystrom (1984) Consideraciones sobre la taxonomia de las especies del genero Cochliodon en Venezuela, Revista UNELLEZ
This paper has a nice distribution map.
C. plecostomoides llanos tributaries of the Orinoco
C. hondae Magdalena system
C. cochliodon Rio Cujaba, Rio Paraguai system Brazil
C. pyrineusi "probably Jamary" meaning the Rio Jamari, afluent of the Rio Madeira?
C. oculeus Florencia, Rio Ortuguesa, afluent of the Rio Caqueta
C. pospisili Rio Palmar, Lake Maracaibo basin
C. taphorni Esequibo

I checked several sources with regards to Eptapterus and can find no records. See especially Fernandez Yepez and Martin (1952) Notas sobre la ictiologica de la region Baruta-el Hatillo.
It lists the following:
Spp restricted to the Tuy; Rhamdia guairensis, Ancistrus brevifilis, Pigidium mondolfi, Chaetostoma dupouii
Spp found in the Tuy
Rhamdia quelen, H. plecostomus, Chaetostoma guairensis, Farlowella acus, Chaetostoma pearsei

I honestly would not know where to start with the Guarico. It is a really big river and I have not seen any detailed studies. Some obvious ones would be H. plecostomoides, Panaque nigrolineatus (it is the holotype location), L 92, L 93, L 193, and that is just loricariids.

Hope those help.
-Shane

[MatsP]

Mats,
I am not sure that these are "bugs" as much as just a lack of available information.

Agreed - but it still means that it should be "as correct as we can make it". And again, I thank you for taking the time to help me here.

1) Listing in the Cat-elog for Chaetostoma sp Rio Zuata is incorrect (not sure why). It is restricted to the Rio Zuata. Rio Zuata connects to the Guarico, Guarico to the Apure, Apure to the Orinoco.

So not the Zuata that joins the Orinoco from the North downstream of Caicara del Orinoco (forming part of the border between Guarico and Anzoategui states). [Darn, why can't those people naming rivers have a register of already used names! ]

I've added it as you've described it, as a tributary to Guarico - it can be changed later if it's incorrect. In fact, this is probably a good time to introduce the "body of water tree-view" to you: Tree view. If you see any river that is wrongly connected, or otherwise "incorrect", let me know.

At this point, I also wish I had a(n even) better map of Venezuela - one with an INDEX of placenames and rivers with coordinates to find them - I'm sure I've spent several hours looking at the map to find a specific river.

2) H. plecostomoides is restricted to the llanos and has never been recorded from the Tuy. H. plecostomus has been recorded from the Tuy (typo?). Check out Lilystrom (1984) Consideraciones sobre la taxonomia de las especies del genero Cochliodon en Venezuela, Revista UNELLEZ
This paper has a nice distribution map.
C. plecostomoides llanos tributaries of the Orinoco

Ok. I'll remove the references to Valencia/Tuy on this one. Just to confuse matters, Fishbase lists this as "Rio Meta Drainage", which of course is part of the llanos tributaries of the Orinoco.

C. hondae Magdalena system

I've got Magdalena and Lake Maracaibo

C. cochliodon Rio Cujaba, Rio Paraguai system Brazil

Occurrence data in Cat-eLog matches this, with a "middle parana" also added

C. pyrineusi "probably Jamary" meaning the Rio Jamari, afluent of the Rio Madeira?

Lists Rio Madeira

C. oculeus Florencia, Rio Ortuguesa, afluent of the Rio Caqueta

Which in turn flows into Japura, which is what I've got in the occurrence records.

C. pospisili Rio Palmar, Lake Maracaibo basin

Not in Cat-eLog

C. taphorni Esequibo

Match

I checked several sources with regards to Eptapterus and can find no records. See especially Fernandez Yepez and Martin (1952) Notas sobre la ictiologica de la region Baruta-el Hatillo.
It lists the following:
Spp restricted to the Tuy; Rhamdia guairensis, Ancistrus brevifilis, Pigidium mondolfi, Chaetostoma dupouii

No Pigidium in fishbase or Cat-eLog, but I found a

Trichomycterus mondolfi

, which is listed as Tuy.

Rhamdia guairensis isn't in the Cat-eLog - Got any pics?

Spp found in the Tuy
Rhamdia quelen, H. plecostomus, Chaetostoma guairensis, Farlowella acus, Chaetostoma pearsei

Aside from C. pearsei that isn't in the Cat-eLog at all, these are now listed under Tuy, along with some others - please feel free to check them over.

I honestly would not know where to start with the Guarico. It is a really big river and I have not seen any detailed studies. Some obvious ones would be H. plecostomoides, Panaque nigrolineatus (it is the holotype location), L 92, L 93, L 193, and that is just loricariids.

I've added H. plecostomoides, P. nigrolineatus, Lasiancistrus tentaculatus (L92), Squaliforma villarsi (L93). Acanthicus sp(L193) is already listed as Orituco, which is a tributary of Guarico, so it should come up on a search of that main river.

I would like to point out that the above species where already listed as "Orinoco" species, which by the definition of the search covers all tributaries of that river. It is fine to list minor rivers on top of larger ones for completeness, but I think I will be very old before ALL occurrences of all existing Cat-eLog entries are covered by ALL the rivers that they can be found in - Ramdia quelen is one of those that would take weeks to list all the rivers that this species can be found in, right?

Hope those help.
-Shane

Yes, thank you very much for the help.

Any more of the same is very welcome - I want this feature of the Cat-eLog to be a useful one, and it would probably be almost impossible for me to gather all the necessary knowledge by myself.

--
Mats

[bronzefry]

Mats, this is already incredibly useful. If we ever get a chance to meet, the beer's on me!
Amanda

[Shane]

Darn, why can't those people naming rivers have a register of already used names!

I could not agree more!

I have a much better understanding now having seen the River Tree. I guess the big question is how far do you want to "split" the systems?

Good maps are ALWAYS hard to get.

I'll have to look and see if I have any Rhamdia pics. If not, someone should be able to supply some.

I checked the Tuy listing and it looks good. It still lists Eptapterus.

H. hondae is listed as occuring in The Magdalena and Maracaibo basins. It is replaced in the Maracaibo Basin by H. popisili. Galvis et al listed C. hondae in the Maracaibo Basin because they considered it a sr. synonym of C. popisili. This has not been upheld, so their records of C. hondae in that drainage are incorrect. This stuff gets complicated!
I would be very careful using any source that contains information from Galvis et al (1997)

Keep in mind that taxonomic changes will also have to be updated in the Distribution since lumping and splitting can radically alter an sp's distribution. Old synonyms all have to be checked as well as they can greatly skew records (as in the case above).

I checked on the Magdalena list. It shows Cheatostoma cf milesi. This fish was caught in the Rio Zuata. Range should be Andean slopes of llanos drainage rivers in Venezuela.

If we ever get a chance to meet, the beer's on me!

Here , here. I'll be in the UK this summer and hope to buy you one as well.

I have a couple of papers that would be very useful to you and will try to get them scanned.

-Shane





[/quote]

[MatsP]

I have a much better understanding now having seen the River Tree. I guess the big question is how far do you want to "split" the systems?

That's a good question. I try to cover at least the rivers that are named on my maps when they are listed in the data I use for the update.

Good maps are ALWAYS hard to get.

Yes, and I guess that the more "undeveloped" the area is, the less likely there is to be a good map. For example, I have great maps of US states in the Delorme Gazzetteer series, which show you really small bodies of waters, and even the smaller roads in towns are there (they may not have NAMES, but they are at least on the map). These maps cover for example Texas in a hundred or so pages, in roughly 17" x 11" format. But South America having an, by average, less affluent population, there's less need of good maps.

I'll have to look and see if I have any Rhamdia pics. If not, someone should be able to supply some.

I checked the Tuy listing and it looks good. It still lists Eptapterus.

Fixed now, got sidetracked last time

H. hondae is listed as occuring in The Magdalena and Maracaibo basins. It is replaced in the Maracaibo Basin by H. popisili. Galvis et al listed C. hondae in the Maracaibo Basin because they considered it a sr. synonym of C. popisili. This has not been upheld, so their records of C. hondae in that drainage are incorrect. This stuff gets complicated!
I would be very careful using any source that contains information from Galvis et al (1997)

Ok, I've removed it

Keep in mind that taxonomic changes will also have to be updated in the Distribution since lumping and splitting can radically alter an sp's distribution. Old synonyms all have to be checked as well as they can greatly skew records (as in the case above).

I checked on the Magdalena list. It shows Cheatostoma cf milesi. This fish was caught in the Rio Zuata. Range should be Andean slopes of llanos drainage rivers in Venezuela.

Ok, will "move" it (it also had Apure on the list - I removed that too, although Zuata ends up in Apure, as per discussion in previous post)

If we ever get a chance to meet, the beer's on me!

Here , here. I'll be in the UK this summer and hope to buy you one as well.

If you fly into Heathrow/Gatwick, I'm not very far (vs. South Africa/America) from there, so let me know when you're here - I'll buy one back, I think I owe you that at least

I have a couple of papers that would be very useful to you and will try to get them scanned.

-Shane

Yes, please


--
Mats

[Jools]

I guess the big question is how far do you want to "split" the systems?

It's all to do with the fishes. If there's a fish restricted to a single puddle in Tanzania, then that puddle goes in the list as its own body of water.

However if all the fishes in the upper reaches of a river are the same as the lower reaches, or a tributary, then there's no need to split.

Jools

[MatsP]

I guess the big question is how far do you want to "split" the systems?

It's all to do with the fishes. If there's a fish restricted to a single puddle in Tanzania, then that puddle goes in the list as its own body of water.

However if all the fishes in the upper reaches of a river are the same as the lower reaches, or a tributary, then there's no need to split.

Jools

For fish in a puddle in Tanzania, thats working fine.

The problem comes, for example, if we search in "Rio Zuata", we should perhaps find a dozen different catfish there, but becasue those fish also appear in the further down-stream Apure, they are not listed as "Rio Zuata".

This is the reverse of the problem that we HAVE solved of searching for occurrences in "Apure" and getting the ones that are in tributaries of that river.

I'm not sure how you would solve this one tho', as it's distinctly different problem (we can't really KNOW that the fish is in the Zuata if it's in Apure, but if it's in Zuata, it's in the Apure drainage (and if that's not what you want, then the "no tributaries" will avoid listing tributary only species).

--
Mats

[Shane]

(we can't really KNOW that the fish is in the Zuata if it's in Apure, but if it's in Zuata, it's in the Apure drainage (and if that's not what you want, then the "no tributaries" will avoid listing tributary only species).

That sounds like a strange statement, but it is 100 percent true. The Rio Zuata (as an example) is a small Andean river that connects to the Guarico when it reaches the plains (llanos). It has a good current and is cooler than the rivers in the plains. Chaetostoma spp present in the Zuata are never found in the Guarico. Same river system, but wrong habitat. Hypostomus plecostomus can be found in both the Guarico and Zuata while Panaque, Acanthicus and other Guarico loricariids are not in the Zuata because it is too fast flowing and cold.
It would be accurate for H. plecostomus to come up under Zuata AND Guarico. Chaetostoma sp af milesi only under Zuata and P. nigrolineatus only under Guarico.
As a general rule, I have found that what I call "generalist" (Hypostomus, Rhamdia, Ancistrus, come to mind) spp usually transcend various environments while "specialists" (say Panaque, Acanthicus, and Chaetostoma) are limited by specific environmental conditions to a well defined area of the drainage in question.
-Shane

[MatsP]

(we can't really KNOW that the fish is in the Zuata if it's in Apure, but if it's in Zuata, it's in the Apure drainage (and if that's not what you want, then the "no tributaries" will avoid listing tributary only species).

That sounds like a strange statement, but it is 100 percent true. The Rio Zuata (as an example) is a small Andean river that connects to the Guarico when it reaches the plains (llanos). It has a good current and is cooler than the rivers in the plains. Chaetostoma spp present in the Zuata are never found in the Guarico. Same river system, but wrong habitat. Hypostomus plecostomus can be found in both the Guarico and Zuata while Panaque, Acanthicus and other Guarico loricariids are not in the Zuata because it is too fast flowing and cold.
It would be accurate for H. plecostomus to come up under Zuata AND Guarico. Chaetostoma sp af milesi only under Zuata and P. nigrolineatus only under Guarico.
As a general rule, I have found that what I call "generalist" (Hypostomus, Rhamdia, Ancistrus, come to mind) spp usually transcend various environments while "specialists" (say Panaque, Acanthicus, and Chaetostoma) are limited by specific environmental conditions to a well defined area of the drainage in question.
-Shane

Yes, and the real problem for me (or anyone else trying to do this) is that you have two choices:
1. Enter every SINGLE river that the species exists in (assuming of course, there is/was/could be a [reliable and accurate] source for this information).
2. Choose a slightly more crude, but for most intents and purposes "good enough" method of describing rivers in a "tree-system", where minor rivers are connected to more major rivers, and be as precise as is reasonable. This means, for example, that Hypostomus plecostomus may only be listed as "Rio Orinoco", which in this case means that it's found in "all of Rio Orinoco and it's tributaries" - this is not entirely accurate, but it's a workable model that works as long as you don't need to know which fishes can be found in exactly Rio Zuata. And whilst that's a nice feature to have.

We (I'm not sure if Jools had any ACTIVE role in this decision, so maybe I should say "I" instead of "we") choose to use the second model.

Unfortunately, computers are rather stupid - hard-working, but don't do anything other than exactly what you told them to do (assuming all parts are workins as intended - if not, there's no telling what they'd do). So we have to work along a model that descibes with sufficient detail which fish can be found where, and the computer will show exactly what has been entered into the data-base. Currently, I reccon we have about 75-80% of south american fish "reasonably correct".

It may be possible, in the future to add a feature like "Populate all tributaries with this species" - that would add all the tributaries of a specific river to the list of occurrences of that species. But that may also not be accurrate (in the above example with the Panaque from Guarico, it doesn't exist in Rio Zuata which flows into Guarico, for example). So this would only be useful for "generalist species". And it would rapidly grow the data-base for the occurrence data - possibly to such an extent that it becomes less useful. [Just to point out that there are over 90 tributataries and sub-sectons (lower/middle/upper) parts to the "Amazon" in the data-base at the moment. That is by far not a complete list of ALL the rivers, just the ones currently in there to describe the species that are currently in the occurrence table].

Anyone interested in looking at it, see the "Tree View" that is linked in the above post.

--
Mats

[Shane]

2. Choose a slightly more crude, but for most intents and purposes "good enough" method of describing rivers in a "tree-system", where minor rivers are connected to more major rivers, and be as precise as is reasonable. This means, for example, that Hypostomus plecostomus may only be listed as "Rio Orinoco", which in this case means that it's found in "all of Rio Orinoco and it's tributaries" - this is not entirely accurate, but it's a workable model that works as long as you don't need to know which fishes can be found in exactly Rio Zuata. And whilst that's a nice feature to have.

I agree you have a real dilemma here. At the end of the day, and far more important than the name of a river or lake, people want to know what conditions their catfishes need to survive in captivity. Secondly, serious biotope folks want to know what fishes naturally occur with what other fishes.

Maybe one possible idea would be to break things down by habitat (e.g. Andean piedmont streams or black water Amazonian tributaries). This would probably be more useful to hobbyists than the name of a little known creek in Brasil. I realize that goes counter to Jools' above statement, but geohydrology is a well defined study (I have many papers) and really tells one more about where a fish lives as far as water conditions than the name of a certain river. What about "habitat" as a subfield?

Just some ideas (thinking aloud on the keyboard).
-Shane

[Shane]

Just as an example to my above statement. South Africa (a good sized country) has 1,000s of creeks and rivers. Add in the other southern African countries and you have possibly tens of thousands.

Break it down by habitats and there are only six in all of southern Africa.

1) Tropical east coast region - The rivers are generally low-gradient mature systems with floodplain reaches.

2) Tropical interior region - These are relatively large, well-watered catchments with a distinct annual flood regime in response to seasonal rainfall cycles.

3) Highveld (temperate) region -

4) Montane-escarpment - Generally high-gradient streams with cool temperatures and low or moderate concentrations of dissolved solids. The fish fauna is of low diversity, but includes characteristic genera such as Chiloglanis and Amphilius.

5) Cape Fold Mountain region - Generally cool-temperate acid water streams. The fish fauna is highly endemic and distinct.

6) Kalahari-Karoo-Namib region - Generally arid with intermittent rivers and temporary pans. Isolated permanent springs and sinkholes. Waters are alkaline and turbid when flowing. The fishes are hardy spp.

Using this sort of system, we have just covered 14 countries with six entries versus 10s of thousands of names of creeks, lakes, rivers, sinkholes, that will not help the average hobbyist understand what their fish need in captivity. We have also avoided the problem of half a dozen "Crocodile Rivers" in southern Africa (or the dozen Rio Blancos in South America).

A nice addition, would be to click on say, "Highveld region" and bring up several "typical" habitat photos of creeks, rivers, substrates, native plants, etc. These would not be linked to the actual spp in the Cat-eLog but a sort of separate "Habitat-eLog" which would mitigate moving this stuff around every time a new taxonomic work is published.

-Shane

[MatsP]

Shane,

I like your idea of "habitat" information. It's certainly something worth considering for some future enhancement of the Cat-eLog.

I think we'd find that there's a few more than 6 worldwide, but certainly nowhere near the 600 odd bodies of water that we've currently got in the occurrence database.

However, I think one doesn't necessarily replace the other, so we need both a "habitat description" and a list of bodies of water to describe which species lives where.

--
Mats

[Shane]

Mats,
I strongly agree that one does not replace the other. "Habitat" info is much more general (macro), but would give a basic idea of water conditions the animal comes from. I see one entry complimenting the other. We could fairly easily collect photos of habitats (e.g. a typical white water Amazon tributary, a typical habitat pic of lake Tanganyika, or a typical highveld stream) than trying to add exact collection locality pics to every Cat-eLog entry. At the end of the day the Rio Orupe, Rio Tamanaco, Rio Tinaquillo, Rio Aguire, and Rio Tinapuy are different rivers, but they are also all Andean piedmont tributaries of the Rio Tinaco and fishes from them require the same captive conditions.
-Shane

[MatsP]

Good, we agree!

I'm not going to add the habitat information right now, but I think it's a neat idea.

To remind ourselves, I'm going to create a new entry in this. The new thread is this one.

--
Mats