re: Shape based assignment as a multivariate measure

Oh man I am going to go way out on a limb here and please don't pillory me for fact please delete this post if necessary based on the fact it probably does not pertain to the subject of the thread because it is not in the strictest sense about "shape based assignment as a multivariate measure" but more so about color and crazy thinking going on in my own mind...

Here we go...

There is one accepted species called homo sapiens. There are three basic types of Homo Sapiens color wise - black, brown, and white SKINNED. Morphology. Biological morphology.

So based upon this example...Homo, white, and brown are the same species right? Does this imply that morphology is not really the best way to differentiate species? I mean does it really come down to the DNA? And an American Indian male is going to carry an unique haplotype code (or something) different from say a northern European or Pacific Islander yet they are NOT different species right? Subspecies?? Do you understand what I am trying to get at?

Is morphology a not so great way to differentiate species really? What about DNA? Is it going to aid taxonomy that much really from what scientist have always done in the past since that Egyptian or Greek guy who started naming everything in some book? So - shape based assignment or color based assignment for that matter as a multivariate measure is not really a valid means of ID'ing species?

A chihuahua, a great dane, a greyhound, and a coyote...they are all wolves right?


re: Shape based assignment as a multivariate measure

Hi Edward,

The question you ask is perfectly valid, so you are justified to asking it, and I think on a topic as complex and contentious as this we should try to address these issues. In general, when it comes to these "morphometric assessments", i.e. measurement of characters, we are trying to actually measure them and then show differences. One of the characters that has been found to be difficult to measure and therefore differentiate is colour, and if you had two biological entities that have everything similar except colour, then you would, in general, not find that they are classified as two species. There may be exceptions, and I stand to be corrected, but this is the general tendency. It is therefore also for this reason that humans have not been classified as differences species, but then this has been a politically hot topic ever since colour differences between humans were noted.

In the context of angels, we see colour differences between scalare and altum, although we also see colour differences between different populations of scalare. Because the colour differences between the scalare cannot be backed up with other significant differences in measurements, they are not viewed as separate species. The bigger differences between altum and scalare are in the measurements that Dave has tried to assess, and which we have actually concluded are not so clear cut and where the problems with regard tothe Rio Negro angels actually lie.

Steve, your point is perfectly valid i.e. that dog x coyote and wolf x coyote would be viewed to be hybrids. What can then happen is that the hybrids cross back repeatedly into the one parent and within three to four generations already, you would hardly see the differences between the hybrids and the original parent. This is also what has been implied in the scientific paper that was quoted before. In the Rio Negro angels we potentially have this problem, but just the measurement of intermediate measurements alone is not enough to prove that they are really hybrids.

Kind regards,

Somerset West, South Africa

Re: re: Shape based assignment as a multivariate measure

Dirk Bellstedt wrote:Hi Dave,

I think that you have now just about exhausted this topic, and I think that we should try to sum up where we stand with all of this with regard to P scalare and P altum and whilst we are about it we should actually include P leopoldi.

Could you perhaps put together a comparison of the morphometric measurements of these taxa including the Rio Negro fishes so that we can say that this is state of affairs as we have it at the moment.
I can put together what has been said in literature, I suppose. I wonder what Heiko's latest works say about counts...
Your extensive discussion of this topic shows us that species concepts or the definition of what we define in this case as an angelfish species is subject to interpretation and that it is not entirely a crisp clear cut scientific assessment. Species definitions are assisted by morphometric measurements, DNA assessments, and even behaviour and habit assessments and last but not least cytogenetic considerations (chromosome counts for the non-scientists amoungst us)(I cannot assess the last thesis because of the fact that I do not speak Portuguese). This also means that as we add new scientific data to what we know we may change our opinions of what constitutes one of the angelfish species.

My personal opinion is that DNA data will be able to give us another set of important insights into how the different angelfish are related and then we will again have to assess where all of this brings us.

Kind regards,

I was thinking that it's a bit more productive to perhaps get photos and find ways to enhance the scales or fins for counting than to worry about what people said before ( although a historical backdrop is nice) .

We have access to so many people with fish and cameras that can capture good images, it seems this area of study can be done much better than has been. We don't need institution-catalogued items in order to find out things :)
that is, it's better to find out by numbers looked at now, than to speculate or even confirm what someone may have said in the past.
For our oncerns, Kullander is basically what we have, I guess.

To do better than previous efforts, we have only to get more than Kullander. Here's # 1 !

Click on Heiko holding bag of altum "full size image". ... &Itemid=45

This study is on topic, I think.
These interpretations of the PC maps have been recently questioned as the original results can be reproduced under models of spatially covarying allele frequencies without any expansion.

In a series of highly influential publications, Cavalli-Sforza and colleagues used principal component (PC) analysis to produce maps depicting how human genetic diversity varies across geographic space. Within Europe, the first axis of variation (PC1) was interpreted as evidence for the demic diffusion model of agriculture, in which farmers expanded from the Near East ∼10,000 years ago and replaced the resident hunter-gatherer populations with little or no interbreeding. These interpretations of the PC maps have been recently questioned as the original results can be reproduced under models of spatially covarying allele frequencies without any expansion.

.... under a broad range of conditions, the gradients in PC1 maps are oriented along a direction perpendicular to the axis of the expansion, rather than along the same axis as the expansion. We propose that this surprising pattern is an outcome of the “allele surfingâ€
You do not have the required permissions to view the files attached to this post.

re: Shape based assignment as a multivariate measure

Isolation plays an extremely important role in species differentiation.

Invasive species are dangerous because they can arrest the long evolving natural balance of an ecosystem and that is bad because it inhibits species differentiation which makes for sterile ecosystems.

Personally I think the DNA on Pterophyllum will tell us that in the beginning there was PERHAPS only the Dumerilii or maybe something very similar...the uplift happened which created two populations of pterophyllum...population a could get to population b but b could not get to a...out of this situation species differentiation occured leading to the rise of altum in the Orinoco and on the "downstream" or below waterfalls side PERHAPS some "allele surfing" led to the rise of the Scalare. Or maybe it is simply different weather and waters made for different fish!

(We really need to explore the concept of mutant genes with regard to pterophyllum wild types *natural forms* in particular especially with all of the talk about red finned red crested this and that...we have a ton of insight into how mutant genes in pterophyllum behave morphalogically wise from the long term selective breeding involving the domestic forms - which are apparently constantly evolving likely due to the mutation of heretofore non expressed but probably ever present heritage genes.)

Back to the topic at hand...I'll have to go back and study it some more but I don't seem to recall a "lump on the forehead" look with altum males in pictures I have seen and I must reiterate I don't know crap about altums other than what I have picked up from some of you folks. Is this a domestic pterophyllum only quality or is it evident in altum, dumerilli, rio negro, etc. as well?

Some silvers from America possibly carrying some wild blood recently in their genome...scales to look at....and spotting...

re: Shape based assignment as a multivariate measure

Hi Edward.
The "humphead" domestic angel males are pretty ugly as they exist. Maybe if it was developed a bit to look like a geophagus it would be interesting. It seems to come along with a big body deformity.

Playing around a bit with GIMP on the angelfish photos in order to see the scales and the patterns, I found this interesting image of concentric cirles of scales which lend the appearance of something of interest in the red dot.

The image also allows a view of the way the fish's topography and general shape influence how the scales are, in number, size, arrangement, and so on. Where you see the light smears diverging, there has to be either more scales or change in scale size. In other words, in many instances a row will "split" into two.
This angel shows a resemblance to a halibut.

I think Heiko's fihs has less scales o the second main bar. That's what I've checked on the altum photos. So far 3 altum seem to have 8 scales, by my count. Heiko's could arguably have seven at most. ... &Itemid=45
You do not have the required permissions to view the files attached to this post.

Re: re: Shape based assignment as a multivariate measure

rag wrote: We have access to so many people with fish and cameras that can capture good images, it seems this area of study can be done much better than has been. We don't need institution-catalogued items in order to find out things :)
that is, it's better to find out by numbers looked at now,
This angel looks like a true altum. The photo is quite clear and enlarging it 300% or more and sharpening,
the dorsal fin rays can be counted. The last three? rays though can not be seen/counted because the fin curved.

Another clear photo of an altum. If i'm not mistaken, one that Heiko caught and photographed. ... /altum.jpg
If we standardize the number of rows beneath the lateral line that we take point "B' to be, on various places on the fish, we can arrive at good comparative figures.
As long as the method is shown and is fully replicable, it's good or better than previous attempts due to easy access to numbers of decent photos.
Clear enough altum photo for counting scales. Enlarging and further enhancing the photo will make the scales more visible.

Re: re: Shape based assignment as a multivariate measure

Hi Dave,

These photographs are really excellent for these measurements, I can see that, so we could continue to count more and more of them. What we actually also need to know is where they have been collected because of the Rio Negro fishes versus "true" Altum.

My question earlier:
Dirk Bellstedt wrote:
Could you perhaps put together a comparison of the morphometric measurements of these taxa including the Rio Negro fishes so that we can say that this is state of affairs as we have it at the moment.
was aimed at establishing a summarizing table of the morphometric measurements for each species and then the RNA as they have been published in the literature. If we had one table it would easier to get the overview. Seeing that you have all the data that the various authors have published, could you perhaps do that for us?

Kind regards,

Somerset West, South Africa

re: Shape based assignment as a multivariate measure

What we could arrive at would be several versions, as format would depend on intended usage.

Although it's not correct to say I have the data ( I only scan articles I find, such as in the Chronicles or recent scientific papers) > I understand from another post here that Jan has the Kullander study.
I did not even have the original Schultz write ups. Now that Fishbase has removed the "Occurrence" chart, it's a bit harder still, to place specimens mentioned as to collection site, without the actual papers ( assuming they supplied all that supplemental data ) to see .

Last year when I tried to sort out what the different authors over time were saying, on the genus and the species, I followed the tracks to say Kullander must be wrong - but Jan corrected me by pointing out Kullander's actual source of leo specimens, rather than how I interpreted the confusing literature ( confusing as seen in Burgess' article) without the Kullander study in front of me.

In any case, I did trace the rest of that historical mix up in the literature all the way through for our most common references.

However, this thread is the first time ( on Burgess ) that I checked the numbers, or cross referenced on the numbers, locales, and number of specimens (and their provenance).

So I believe that Kullander is most likely correct, in what he looked at.
However, there arise the problems and questions explored here:

1. Did he examine 3 fish 30 fish or 300 ? How many locales involved ?

2/ Is his meaurement to be taken as "maximum length", "standard" (whatever that means ) length, or "mean average" length ?

I'm finding that the altum photos are all coming in with high scale counts
and I am tending to think that Kullander is correct in "end results" - and that scale counts are an effective means of identifying altum assuming none turn up that have under 46 scales.

Dirk, this is another occasion where I think that method needs to be examined. We do not need to have complete collection locale information, to do some things.

As I explained, this is mostly "loosey goosey" stuff anyway. If one decides what must be Pterophyllum altum DNA without ever having a reference DNA, one has nothing but a difference showing.

This is not a whit better than putting in 50 or 100 altum barcodes from our own fish. We could say we found difference, fish to fish or even locale to locale but we can't make very strong inferences. At least we would use a good reference specimen or two.

My point is that our own eyeballs tell us which fish ( put in negative terms) are not altum, and which might not be altum, but are also not scalare.

If we note which altums look a bit different from other types, and they SHOW a DNA difference, then it doesn't matter ! ( as we have found a difference that then can be investigated further by us - because we know location for SOME of the altum ( the most important, the type reference locales included ) - and then can be further investigated by a fuller study checking for difference at the possible locales or phenotypes brought into question.

What we would then have located is a differing phenome that matches a differing DNA. Possibly we might find a differing DNA but no difference in appearance.

This is surely as strong investigation as we have already noted being done. Even more important is true modesty in realization of limitations for claims.

Kullander's university site publishes it as "maximun length". Natasha uses "Standard Length" (with the same 6.5 and 7.5 cm figures applied ) and it's stated as taken from Kullander.

If this was an occasion where there was only one species in the genus it would not be as critical as when species differentiation is made on these figures.

Whereas those researches are not explicitly explained ( method, terms, and real data ) We are not privy to kowing what parts of the Kullander "pointers" Natasha used - was standard length a factor ? Or did she depend on scale counts ? As to data..lacking, and not one photograph of the subjects of interest...

...our method of eyeballing, together with exporter/importer/collector info, added to scientific knowledge from the literature we know about, allows us a more comprehensive answer and a more expertly guided answer. Photo from Snookn on aquabid shows fish he says are Nanay, and we know we can ID. these fish.
We need only image google and find such a fish with locale given, fresh out of the river

This is light years more credible research than Axelrod could ever provide. How can Science justify holding out on realty ?

Modern times: relatively huge amounts of VERIFIABLE HD photographic evidences and multiple first hand reports ( some even with GPS reports ) available to those with interest.

This of coure was never available to us in the days of books and papers that nobody could ever see unless they subscribed to many journals.

That situation is in the past, has been for a few years now.

I see the version for the forum being for hobbyist interest and the state of knowledge would be acceptance of Kullander's analysis for the most part, plus study of a good number of photographs giving us amounts of real data confirming what might otherwise be considered a well-categorized summary, but perhaps lacking in amount of data points.

If scale count if indeed so different ( no overlap), that's a good proxy.

Addendum 1:
I understand from another article, that Kullander had examined scalare from Brazil, Peru and Guyana, but the article made no mention of numbers.

First pic from Aquabid. from "snookn", posting ad for Peruvian Nanay.

We are capable of obtaining an ID on this fish from online photos of fish direct from river as seen below. For the first step of such a 3 part phylogenetic study as Natasha's, this kind of evidence for the kind of fish being examined, would have been far better evidence than 500 mushy words offering zero description or "results of measuring", for any specimens

Addendum 2: On consideration, I've assumed that the "standard" in "standard length" is probably refering to the approach of using snout to tail length. This may offer help in understandnig the term , but offers no help, however, in determining if the length is an average, or if it's a maximum. We'd know they measured them in the same way, that's all.

You do not have the required permissions to view the files attached to this post.

re: Shape based assignment as a multivariate measure

A simplified opinion on the matter.

The Tapajos and St Isabel, I believe are two distinct types of Pterophyllum. The only thing I see they have in common other than being angelfish is that they both show an interesting amount of red on the head area. Tapajos additionally shows a lot of red into the fins but it has a all the traits of P. scalare otherwise. Bar width, bar coloration (black vs. the typical altum dark reddish-brown), head shape, scale size and numbers, etc, all are closer to P. scalare in Tapajos. It doesn't look nothing like an altum to me.

St Isabel has a definite influence of P.altum. Bar width, coloration and all other traits, and has the interesting concentration of red on the head. It is an altum with a red head. I won't go into the P.altum or P.scalare discussion here, whether St Isabel is a true P.altum or not... I suppose not... but that's one question we should be able to answer not far from now, hopefully.

Now, a respected European OTF importer mistakenly labels his fish based on info that I suppose he considers accurate (coming from the exporter) and we end up with a big mess.

The photos referred by Ken in above links (Heiko and the other pic) are true P.altum, no doubt, but they are young specimens, far from mature adults. Altums change considerably once they approach maturity. Edward's reference to the "classic" mature altum is pretty much exact.

God listens. He may not always give you everything you want, when you want it; but he always gives you what you really need, when you need it most!