How many avian orders are there? (3 Viewers)

[Jim M.]

At risk of escalating on a thread that already seems too argumentative for my liking - here is a blog that features several examples of nighthawks perching in all sorts of positions: http://backyardbirdinginsoutheastte...huac-national-wildlife-refuge-common.html?m=1

There are several other examples online, although this one is nice because there are several all on one page. But regardless, none of these are very comparable to the vertical clinging exhibited by most swifts.

Also, the "cigar-shaped body" so often referred to in field guides is almost invariably used to describe the tubular and blunted-at-both-ends body type of a swift such as Vaux's, Chimney, etc. Nighthawks are "front heavy" in body shape and taper or attenuate to a long, not blunt, tail. I'll again state that I've never before heard this referred to as shaped like any sort of cigar, nor like a swift. In fact, out of all the students and birders I've dealt with (mostly eastern U.S.), I've run into exactly zero instances of confusing swifts and nighthawks - although I'm certain it has happened somewhere.

And now that I've said all this, I'm not replying any more on this matter. I hope its obvious that I'm not looking to pick a fight here, but if I came across that way, I do apologize.

I'm not directing this statement toward Jim, but mostly to everyone contributing here - some of the comments/arguments on this thread are starting to look a little snarky to an outside perspective - let's remember to be courteous if we can.

Thanks,
Kirk

Kirk,

My original post was directed at Niels and was merely trying to massage his intuitions by suggesting there were broad analogies between some members of the traditional caprimulgiformes and the traditional apodiformes, such that combining the two orders wasn't so crazy because some members of each seemed more similar to each other than to birds in other orders. Since I was comparing birds that we all agree are at least in different families, the analogies were of course broad and subject to interpretation. Trying to nitpick the analogies was really missing the point. To the extent you wanted to break down the analogies, I think the way to do it would be to try to come up with species in other orders that are just as, or more similar--e.g. swallows would be one candidate.

As for your specific claims, the very page you link to confirms what I have been saying about nighthawks. Beneath the single photo that may show a nighthawk perching perpindicular to a wire (I say may because the angle makes it a bit ambiguous) it says:

One even perched on the barbed wire. This is an unusual pose because the bird's legs are short and its feet small in relation to body size. It's not really made for perching. When it perches in a tree, often it will sit lengthwise on the limb rather than across the limb as most birds do.
​

Your comment that swifts are even more limited is again missing the point. The point is they are both quite limited--unlike birds in other orders such as the passeriformes, coraciformes, coliformes (mousebirds), etc.

As for confusing swifts and nighthawks, I have personally confused them for an instant when they are flying together and viewed at considerable distance.

And if you don't like "cigar-shaped" for nighthawks, then "narrow tubular body" will do just as well to make my point. This is in contrast to, e.g. a crow or Belted Kingfisher, for example.

 

[Mysticete]

Sorry for sounding sometimes very official. Just in case anybody wondered - I don't make my money from birds either.



Always time to learn something new. So, please, answer the question: having the biggest data set available, which is the best division of birds and of mammals into groups? The best is the one, traditional, where similarity inside groups is biggest and similarity between groups is smallest. And how much it is better from the second-best group? It should be very easy question, since you say it is 50 years going on.

I don't follow phylogenetics closely, but in casual glance it is occupied more with the question 'what is related to what and not to what'. The question is different: 'a what level the groups are most clearly visible'.

That's because what you want doesn't exist. Evolution, by it's very nature, involves incremental changes in morphology. The Linnean ranking system was designed long before the framework of the Theory of Evolution was laid out.

If you tried this approach with morphology, you would face many shortfalls. First, morphological characters by their very nature have a degree of arbitrariness, since you are delimiting continuous features into discrete units. And people disagree on the usefulness of certain characters and so analyses would be based on any number of differently sized character datasets. So any classification you formulated would almost certainly vary based on the datasets used, which would not be agreed on. Secondly, there really isn't any way to accommodate fossils, which possess intermediate features, into such an arrangement. Some of these would be "equally" similar to different related groups, so what would we do? Focusing on a narrow set of characters for delimitation makes thing even more arbitrary

Molecular data has its own pitfalls, since genetic divergence is at least in part going to be related to factors such as population and generation size.

I said that this method already existed in my reply, because what you want is pretty much phylogenetics. using shared features and differences to group taxa. But any cut off point you define in that branching structure is inherently arbitrary, and merely a result of "tradition" or some other criteria you have chosen, whether it's a set of shared features you think are especially important or when the initial divergence evolved.

The "solution" to the linnean rank problem is simple. Don't worry about it. Name new clades as needed but don't obsess over whether they rank as a family, supraorder, infraorder, or any of the other units that have been added over the years as biologists have tried to reconcile the problem of using evolutionary trees with a system of classification invented before the concept.

 

[fugl]

. . .The "solution" to the linnean rank problem is simple. Don't worry about it. Name new clades as needed but don't obsess over whether they rank as a family, supraorder, infraorder, or any of the other units that have been added over the years as biologists have tried to reconcile the problem of using evolutionary trees with a system of classification invented before the concept.

Indeed. How long will it be, I wonder, as the evolutionary tree continues to fill out, before such labels devolve into mere folk terms like "pachyderms" or "raptors"?

 

[jurek]

That's because what you want doesn't exist. Evolution, by it's very nature, involves incremental changes in morphology.

This is no argument that groups don't exist. There are lots of clearly separated groups around, because intermediate forms died out.

If you tried this approach with morphology, you would face many shortfalls. First, morphological characters by their very nature have a degree of arbitrariness

Well, people who study fossil animals have only morphological characters, and would be very surprised to hear that the idea is wrong. Now you started criticising the idea of matrix of characters, which is a starting point to phylogenetic trees, too.

But any cut off point you define in that branching structure is inherently arbitrary, and merely a result of "tradition" or some other criteria you have chosen

I am afraid you don't understand: the best cutoff is not necessary arbitrary. You can find the best cutoff using maths. Which was the point of my original post.

The "solution" to the linnean rank problem is simple. Don't worry about it.

Until somebody actually looks for the best cutoff, and finds there isn't one, this is only your opinion.

My opinion is quite different. The Linnaean ranking of genus, family, order holds for several centuries, and both scientists and laymen tend to independently agree what belongs to many well defined groups (e.g. what is a bird, a parrot or a bear). That the system holds, and people independently arrive into it, suggests that the best division of biodiversity into groups might exist objectively, and might partially match the Linnaean ranking.

 

[jurek]

Indeed. How long will it be, I wonder, as the evolutionary tree continues to fill out, before such labels devolve into mere folk terms like "pachyderms" or "raptors"?

It already ceases to exist. I don't recall any publication seriously putting whales between pigs and the rest of even-toed ungulates (the infamous group Cetartiodactyla). If you look into conservation groups like IUCN, they use what you call folk terms (e.g. African Elephant and Rhino protection working group = pachyderms, Raptor working group, eagle conservation committee).

 

[fugl]

It already ceases to exist. I don't recall any publication seriously putting whales between pigs and the rest of even-toed ungulates (the infamous group Cetartiodactyla). If you look into conservation groups like IUCN, they use what you call folk terms (e.g. African Elephant and Rhino protection working group = pachyderms, Raptor working group, eagle conservation committee).

I'm not sure I follow what you're trying to say here. . ..

 

[Mysticete]

This is no argument that groups don't exist. There are lots of clearly separated groups around, because intermediate forms died out.

A taxonomy that excludes 99% of all life that has existed (or one that ignores well supported phylogenetic relationships) is not a taxonomy that is going to gain any traction in science

Well, people who study fossil animals have only morphological characters, and would be very surprised to hear that the idea is wrong. Now you started criticising the idea of matrix of characters, which is a starting point to phylogenetic trees, too.

You don't think said people are aware of this. I write this as someone who spent almost all of Sunday character coding a character matrix of extinct seals. Pretty much almost all of us realize that issues are present in morphological data. hopefully those issues are minimized by including as many characters, taxa, and specimens as possible

That said while we may use a certain morphological complex to designate a group, we don't then try to use those same characters to define all group

I am afraid you don't understand: the best cutoff is not necessary arbitrary. You can find the best cutoff using maths. Which was the point of my original post.

Until somebody actually looks for the best cutoff, and finds there isn't one, this is only your opinion.

The computer and its math might not be arbitrary, but the computer still needs to have data collected and entered into it. What data that is selected and entered in, and how it is entered in, is what is arbitrary

My opinion is quite different. The Linnaean ranking of genus, family, order holds for several centuries, and both scientists and laymen tend to independently agree what belongs to many well defined groups (e.g. what is a bird, a parrot or a bear). That the system holds, and people independently arrive into it, suggests that the best division of biodiversity into groups might exist objectively, and might partially match the Linnaean ranking.

That people believe something for hundreds of years doesn't mean it's automatically correct or efficient...even a casual glance at world history reveals many long held ideas about science have proven to be unuseful and later jettisoned.

 

[Mysticete]

It already ceases to exist. I don't recall any publication seriously putting whales between pigs and the rest of even-toed ungulates (the infamous group Cetartiodactyla). If you look into conservation groups like IUCN, they use what you call folk terms (e.g. African Elephant and Rhino protection working group = pachyderms, Raptor working group, eagle conservation committee).

Both wikipedia and the Marine Mammal Society's official checklist accept Artiodactyla/Cetartiodactyla as containing whales, and at this point I think it's accepted knowledge by the scientific community concerned with this. That governmental bodies don't change higher level taxonomy is irrelevant, since whether something is an order or a family seldom influences species management. And no one is saying you can't use raptor or pachyderm. They are still useful terms, just not useful terms that have taxonomic definition.

 

[jurek]

I'm not sure I follow what you're trying to say here. . ..

That modern taxonomic groups are useless for practical purposes, therefore there is shift back to using groups which are not monophyletic (or even not evolutionary right) in all disciplines of biology outside narrowest discussion about taxonomy: birdwatching, conservation, fisheries etc.

Typical example is HMW, which has volumes "carnivores" "hoofed mammals" and "sea mammals", instead of for example "cetartiodactyls" or "land and water carnivores".

 

[jurek]

A taxonomy that excludes 99% of all life that has existed (or one that ignores well supported phylogenetic relationships) is not a taxonomy that is going to gain any traction in science

You don't understand. Objective groups are equally well findable in extinct and modern groups. I think, you are worried that thee is always one ancestor linking groups. This is not a big problem. Ancestor may be closer to one of the groups, and in most practical cases is simply unknown. In the worst case, there is only one un-assignable species. The same problem exists, actually, with phylogenetic trees.

That people believe something for hundreds of years doesn't mean it's automatically correct or efficient...even a casual glance at world history reveals many long held ideas about science have proven to be unuseful and later jettisoned.

You don't understand or pretend argument from authority. I propose that the fact that people independently arrive at the same groups of animals suggests that these groups are objective entities, even if people don't define them formally. The real test would be actual application of cluster analysis or similar group-finding method.

Don't I notice, by any chance, annoyance that cluster analysis may make lots of taxonomic work irrelevant? Similar to annoyance of classical biologists when taxonomy based on DNA appeared?

 

[D Halas]

You don't understand or pretend argument from authority. I propose that the fact that people independently arrive at the same groups of animals suggests that these groups are objective entities, even if people don't define them formally. The real test would be actual application of cluster analysis or similar group-finding method.

Don't I notice, by any chance, annoyance that cluster analysis may make lots of taxonomic work irrelevant? Similar to annoyance of classical biologists when taxonomy based on DNA appeared?

This has been done, over 50 years ago. It was called phenetics; you can read all about it in "Principles of Numerical Taxonomy" by Sokal and Sneath. It fell out of favour largely because the groups produced by phenetic methods aren't nearly as useful as those we discover by using cladistic methods.

 

[Mysticete]

That modern taxonomic groups are useless for practical purposes, therefore there is shift back to using groups which are not monophyletic (or even not evolutionary right) in all disciplines of biology outside narrowest discussion about taxonomy: birdwatching, conservation, fisheries etc.

Typical example is HMW, which has volumes "carnivores" "hoofed mammals" and "sea mammals", instead of for example "cetartiodactyls" or "land and water carnivores".

You mistake people using common terms with some sort of actual movement to "classify" animals by specific terms. A casual read of the Carnivore volume would actually prove your point very mistaken. Otherwise Families like Eupleridae or Mephitidae wouldn't have been recognized. They are still using an up to date taxonomy for there organization. They just move stuff around, because people are more likely to buy an entire book on marine mammals, etc than 3 rodent volumes.

 

[Mysticete]

You don't understand. Objective groups are equally well findable in extinct and modern groups. I think, you are worried that thee is always one ancestor linking groups. This is not a big problem. Ancestor may be closer to one of the groups, and in most practical cases is simply unknown. In the worst case, there is only one un-assignable species. The same problem exists, actually, with phylogenetic trees.

I'd love for you to walk me through how you objectively determine family and higher relationships for whales and their close relatives. At the base you have animals superficially similar to chevrotains, than vaguely wolf like pakicetids, followed by seal-esq protocetids, and finally basilosaurids, which at least actually start to look like the common conception of "whale." With phylogenetic trees, we can at least place them in the tree as successive stem taxa, and name the clades without worrying about rank.

You don't understand or pretend argument from authority. I propose that the fact that people independently arrive at the same groups of animals suggests that these groups are objective entities, even if people don't define them formally. The real test would be actual application of cluster analysis or similar group-finding method.

Don't I notice, by any chance, annoyance that cluster analysis may make lots of taxonomic work irrelevant? Similar to annoyance of classical biologists when taxonomy based on DNA appeared?

Frustration more than annoyance that I am clearly not getting my point across.

That people independently arrive at the "same" conclusion means little, because "common sense" is often wrong. Again a cursory view of the history of science identifies many casual observations as being wrong.

Pachydermata is a great example. It was a group that basically was designated to include "large hairless thick-skinned herbivorous mammals". But that is practically it. All three groups have very different patterns of evolutionary history, life history, and anatomy. Grouping them together does more hindrance than harm.

 

[Nutcracker]

Pachydermata .... Grouping them together does more hindrance than harm.

In many respects yes, but equally, they are all most severely threatened by the same factor (poaching), so very similar conservation strategies apply

 

[Mysticete]

In many respects yes, but equally, they are all most severely threatened by the same factor (poaching), so very similar conservation strategies apply

Poaching is going to group a lot of very different animals together. From a conservation viewpoint, the reproductive biology of these animals is all very different; while elephants and hippos don't seem to be too adverse to captive breeding (if you have the room), most rhinos are far more challenging.

Also, Elephants are very social animals that require a very different social upbringing, versus more solitary Rhinos. Elephants that lose their matriarchs or older members of the tribe often lose access to knowledge of important food and water resources. It also means some elephants growing up do not receive enough socialization. Which in turn leads to "delinquent" male elephants murdering the hell out of rhinos (one recent South African park lost 36 rhinos to elephants)

So as you can see ...very strong differences in life history that separate these groups does actually influence the conservation and management of endangered populations.

 

[jurek]

This has been done, over 50 years ago. It was called phenetics; you can read all about it in "Principles of Numerical Taxonomy" by Sokal and Sneath. It fell out of favour largely because the groups produced by phenetic methods aren't nearly as useful as those we discover by using cladistic methods.

Well, no wonder, with no molecular data and with only hand calculators! It would be interesting to try it again, given that groups like grebe-and-flamingo clade or artiodactyl-and-whale clade are about as useless as it can be.

I was not actually talking about phenetics, but using clustering and machine learning to objectively delimite groups.

PS. Just read quickly - phenetics, contrary to what you written, is well and alive. Whoever does genome comparison, uses phenetics more than cladistics - ACGT nucleotides are equally good characters, looking at the genome level it is hard to determine derived nucleotide sequences.

 

[jurek]

I'd love for you to walk me through how you objectively determine family and higher relationships for whales and their close relatives.

The same way how you determine phylogenetic tree, only using different algorithm - SVMs or silhouette followed by k-means clustering, instead of hierarchical clustering.

You seem not to understand that there are many equally valid mathematical methods as hierarchical modelling. They all use the same original data, only the focus is on finding and delimiting groups, not one-to-one neighbors.

They all partially match each other - for example, where you have very long alone branch of a phylogenetic tree (e.g. Kagu to other birds), there will be a very clearly demarcated border of a cluster.

A year ago I would probably run the calcularion if you actually send me the data matrix of you whales. Now I have too much time pressures.

Pachydermata is a great example. It was a group that basically was designated to include "large hairless thick-skinned herbivorous mammals". But that is practically it. All three groups have very different patterns of evolutionary history, life history, and anatomy. Grouping them together does more hindrance than harm.

Quite the opposite. For ecologists and conservationists, this is very useful group. Only they call it megaherbivores. See eg. the classic book:
http://www.cambridge.org/ch/academi...ivores-influence-very-large-body-size-ecology

Elephants, rhinos and hippos have more in common for an ecologist than eg. elephants and hyrax, hippos and warthogs or rhinos and zebras. All pachyderms are too big for predators as adults, all are unseasonal breeders due to very long gestation, are strong enough to be landscape engineers by destroying vegetation. All disperse seeds of many plants which are too big to other animals.

For a conservationist, all pachyderms are threatened primarily by poaching, all require large reserves, all have very low birthrate, all can support ecotourism and be good ambassador species.

 

[D Halas]

Well, no wonder, with no molecular data and with only hand calculators! It would be interesting to try it again, given that groups like grebe-and-flamingo clade or artiodactyl-and-whale clade are about as useless as it can be.

I was not actually talking about phenetics, but using clustering and machine learning to objectively delimite groups.

PS. Just read quickly - phenetics, contrary to what you written, is well and alive. Whoever does genome comparison, uses phenetics more than cladistics - ACGT nucleotides are equally good characters, looking at the genome level it is hard to determine derived nucleotide sequences.

It doesn't matter what kind of data are used, and the calculations back then, even using only hand calculators, were nonetheless exact, given small enough sets of taxa. Molecular data and more taxa won't change the fact that cladistic groupings are more useful for biological inference than any other kind of grouping. See Farris 1979, "The Information Content of the Phylogenetic System":

https://www.jstor.org/stable/2412562?seq=1#page_scan_tab_contents


And no, no one does phenetic analysis any more –apart from neighbour joining trees, which no one would ever use who actually wants a reliable reconstruction of the relationships within a group. Any phylogenetic tree you'll find in any reputable journal, whether a Bayesian, maximum likelihood, or maximum parsimony tree, implicitly uses cladistic principles of outgroup analysis to determine the polarisation of characters – and yes, that includes DNA sequence data.

 

[jurek]

I see here a paradigm waiting to be overturned. Mostly, because non-taxonomists want labels, easy to use division into clear and objective groups. Like the first question of this thread.

 

[Mysticete]

non taxonomists already have their labels...they are called "common names". There is no need for common and taxonomic names to be the same thing.