Time calibration and Linnean ranks in birds (3 Viewers)

[Snapdragyn]

My personal list uses such an approach. I chose dates with the aim of minimizing taxonomic while connecting date thresholds to geological time periods. I won't say this was the 'best' approach, but it was simply very easy to find dates to consider that way.

With orders set at roughly* 66 MYA, I end up w/ 45 orders.
With families set at roughly* 23 MYA, I end up w/ 603 families.

Gotta head to work now, but if anyone is interested I can do a similar 'change by group' list later. I know most of the family increase on my list comes in the non-passerines; there are a LOT of clades there far older than some of the ridiculous (IMO) 9-primaried oscine 'families' (some of which are only tribes in my list).

* I took a somewhat different approach than you on the 'soft/hard limit' issue. For me, anything below <rank threshold> could NOT be that rank. Anything between the threshold & 1/4 of the time to the threshold above could be the rank, but didn't have to be. If, for example, there were multiple branches in a clade all falling within that 'soft zone', I might split the 1st branch at the relevant rank while demoting the next branch to the next lower rank (especially if there otherwise would be no splits at that next lower rank). I felt this fuzzy approach at times allowed for a better reflection of clade branching within the list.

 

[Acanthis]

European GP is breeding in the UK and Denmark it seems, not exactly places I would consider Tundra ...

European Golden-Plover - Abundance Map - eBird Status and Trends

Relative abundance is depicted for each season along a color gradient from a light color indicating lower relative abundance to a dark color indicating a higher relative abundance. Relative abundance is the estimated average count of individuals detected by an eBirder during a 1 hour, 2...

science.ebird.org

Well I can't really speak for Denmark or the rest of the UK but here in Scotland I've only ever seen breeding golden plovers in cold, bleak, windswept, upland places. Not true tundra perhaps but as close as you can get to it on the island of Britain.
Also, a full quarter of the group are not 'golden' 🤣

 

[Acanthis]

Title of book on them?

Cheers Mike!
That'll be it

 

[Acanthis]

My personal list uses such an approach. I chose dates with the aim of minimizing taxonomic while connecting date thresholds to geological time periods. I won't say this was the 'best' approach, but it was simply very easy to find dates to consider that way.

With orders set at roughly* 66 MYA, I end up w/ 45 orders.
With families set at roughly* 23 MYA, I end up w/ 603 families.

Gotta head to work now, but if anyone is interested I can do a similar 'change by group' list later. I know most of the family increase on my list comes in the non-passerines; there are a LOT of clades there far older than some of the ridiculous (IMO) 9-primaried oscine 'families' (some of which are only tribes in my list).

* I took a somewhat different approach than you on the 'soft/hard limit' issue. For me, anything below <rank threshold> could NOT be that rank. Anything between the threshold & 1/4 of the time to the threshold above could be the rank, but didn't have to be. If, for example, there were multiple branches in a clade all falling within that 'soft zone', I might split the 1st branch at the relevant rank while demoting the next branch to the next lower rank (especially if there otherwise would be no splits at that next lower rank). I felt this fuzzy approach at times allowed for a better reflection of clade branching within the list.

That sounds fascinating!
I'd love to compare it to my own little 'collapsed hierarchy' mind-excercise👍

 

[Mysticete]

My personal list uses such an approach. I chose dates with the aim of minimizing taxonomic while connecting date thresholds to geological time periods. I won't say this was the 'best' approach, but it was simply very easy to find dates to consider that way.

With orders set at roughly* 66 MYA, I end up w/ 45 orders.
With families set at roughly* 23 MYA, I end up w/ 603 families.

Gotta head to work now, but if anyone is interested I can do a similar 'change by group' list later. I know most of the family increase on my list comes in the non-passerines; there are a LOT of clades there far older than some of the ridiculous (IMO) 9-primaried oscine 'families' (some of which are only tribes in my list).

* I took a somewhat different approach than you on the 'soft/hard limit' issue. For me, anything below <rank threshold> could NOT be that rank. Anything between the threshold & 1/4 of the time to the threshold above could be the rank, but didn't have to be. If, for example, there were multiple branches in a clade all falling within that 'soft zone', I might split the 1st branch at the relevant rank while demoting the next branch to the next lower rank (especially if there otherwise would be no splits at that next lower rank). I felt this fuzzy approach at times allowed for a better reflection of clade branching within the list.

As a paleontologist, I am very skeptical of the idea that most bird orders were present at or around the time of the extinction of the dinosaurs, So I would personally suggest instead to use the PETM event at 56 million years ago. This seems to fit with studies suggesting younger origination events and lines up with explosive diversification in mammals as well. I would guess that you probably had one or two paleognath birds (probably the distant ancestor of the ostrich and the ancestor of everything else), one stem galliform, one stem anseriform, and ancestral Neoavian.

Please feel free to list your classification here! It provides moral support for us with weirder ideas of what should be split or lumped!

 

[njlarsen]

Well I can't really speak for Denmark or the rest of the UK but here in Scotland I've only ever seen breeding golden plovers in cold, bleak, windswept, upland places. Not true tundra perhaps but as close as you can get to it on the island of Britain.
Also, a full quarter of the group are not 'golden' 🤣

Definition of Tundra can be found on the page linked to below. An excerpt:

The temperatures are so cold that there is a layer of permanently frozen ground below the surface, called permafrost. This permafrost is a defining characteristic of the tundra biome.

Tundra: Mission: Biomes

climate change, global climate change, global warming, natural hazards, Earth, environment, remote sensing, atmosphere, land processes, oceans, volcanoes, land cover, Earth science data, NASA, environmental processes, Blue Marble, global maps

earthobservatory.nasa.gov

And the juvenile has yellowish spots even if these quickly fade to whitish
Niels

 

[Snapdragyn]

I'm not at all informed on the paleontology, but I've seen multiple datings that suggest most or even all regularly recognized modern bird orders crossed the dino extinction event.

I even chose the date originally specifically because it was the oldest date I could find at the time (I started this sometime in 2010) that would still retain a singular Passeriformes (although I've recently seen a paper suggesting that the split of Acanthasittidae may even be older than that).

OK, for the thought exercise purposes, here goes. Only extant families & orders included.

Struthioniformes - 1 family
Rheiformes - 15 families
Anhimiformes - 1 family
Anseranatiformes - 1 family
Anseriformes - 7 families
Megapodiiformes - 6 families
Galliformes - 20 families
Gruiformes - 16 families
Charadriiformes - 18 families
Scolopaciformes - 39 families
Musophagiformes - 3 families
Otidiformes - 8 families
Mesitornithiformes - 2 families
Pterocliformes - 5 families
Columbiformes - 29 families
Cuculiformes - 22 families
Caprimulgiformes - 8 families
Steatornithiformes - 1 family
Nyctibiiformes - 4 families
Podargiformes - 5 families
Apodiformes - 19 families
Phoenicopteriformes - 4 families
Opisthocomiformes - 1 family
Phaethonhtiformes - 1 family
Eurypygiformes - 2 families
Gaviiformes - 2 families
Ciconiiformes - 3 families
Suliformes - 4 families
Threskiornithiformes - 9 families
Pelecaniformes - 18 families
Sphenisciformes - 1 family
Procellariiformes - 17 families
Coliiformes - 4 families
Leptosomiformes - 1 family
Trogoniformes - 6 families
Bucerotiformes - 9 families
Coraciiformes - 11 families
Piciformes - 18 families
Strigiformes - 16 families
Cathartiformes - 1 family
Accipitriformes - 20 families
Cariamiformes - 1 family
Falconiformes - 3 families
Psittaciformes - 22 families
Passeriformes - 194 families

I think the main point of interest in the exercise is how 'off' current taxonomy might be considered in the gross discrepancies of how it treats different groups. Even within Passeriformes, there are areas where I have multiple family-level splits (antbirds come out to 7 families under this system) while others groups are much lumped by the same standard (only 4 families of 9-primaried oscines). To what extent the heavy splitting of non-passerine groups here is an artifact of applying a single standard to longer-lived groups I do not know, but it does at least lead me to ponder if a system that properly adjusted for evolutionary rate would still lead to the large families we currently see or instead suggest making at least some of these splits.

 

[Jim LeNomenclatoriste]

for my part, I would like to use a classification which combines molecular data and the structure of egg shells as studied by Konstantin Mikhailov

 

[jurek]

You lump together taxa very dissimilar biologically and ecologically, and split much more similar taxa into two or more very similar groups. Basically, you have shown that time-calibrated, monophyletic groups are an impractical method for classification of birds.

It fits well the current understanding that evolution proceeds at different speeds depending from the strength of the natural selection. In this context, interesting is that birds living in the ecologically stable rainforest environment produce many deep genetic divergences similar ecologically (forest partridges, forest rails, forest falcons, nightjar relatives), while birds of unstable open habitats and waterbodies produced relatively many dissimilar but relatively closely related groups.

I feel you could create a classification of birds which is more practically usable and more likely to be adopted if you changed the approach. Define groups which are monophyletic in a narrow sense of having a common ancestor and including some descendants, and defined by morphological and ecological similarity. This method is objective and replicable. It requires a list of morphological characters and finding groups most similar within each other but most dissimilar between each other. It can include extinct species which genetic information is not accessible. It will also produce groups which are practically useful, not ones linking species which have visually, biologically and ecologically little in common (likes of the honeyguide-woodpecker group, flamingo-grebe group etc.).

 

[jurek]

I would be happy with lumping some of the oscines. Enlarged Corvidae and "one-Tyrannidae" were in my very first Sibley & Monroe-based list. I agree that generation times should play a role, but currently the oscines are ridiculously oversplit.

I fully agree!

The general rule is that during classification, obtaining many alternative results is an indication that the group is uniform and should not be subdivided. When every attempt of defining subgroups results in different ones, it is an indication that no real subgroups exist - the group is objectively uniform. This is very much true for passerines.

BTW - the premise of Linnaean classification comes from the time when evolution was not yet known and biology and paleontology were only superficially understood. It tried to classify plants and animals like human society, using the concept of relatively few hierarchical steps with equal members - and using names from the human society like 'family' 'tribe' 'kingdom'. We currently know that evolution does not work like that, especially there are no pulses and lulls of evolution which would produce hierarchical steps of diversity. Only rarely, and not uniformly among different groups, there are ecological radiations which can really produce many taxa of equal rank.

 

[Xenospiza]

And the juvenile has yellowish spots even if these quickly fade to whitish

Grey/Black-bellied Plover is called "Silver Plover" in Dutch, but a really old name is the odd "Golden Lapwing".
I think Golden Plovers (no hyphens please) is a sensible name; Black-bellied Plovers would work as well.

 

[jurek]

Large plovers is also possible.

 

[Acanthis]

Grey/Black-bellied Plover is called "Silver Plover" in Dutch, but a really old name is the odd "Golden Lapwing".
I think Golden Plovers (no hyphens please) is a sensible name; Black-bellied Plovers would work as well.

I love "Silver Plover". Nice concept which fits well with impressions the summer plumage lends.
Golden lapwing is just weird🙃

 

[Snapdragyn]

You lump together taxa very dissimilar biologically and ecologically, and split much more similar taxa into two or more very similar groups. Basically, you have shown that time-calibrated, monophyletic groups are an impractical method for classification of birds."

On the contrary, I would argue this shows that using objective criteria produces different results than using subjective criteria such as 'similarity' (to the human eye). Which criteria are preferred is another question entirely.

 

[Acanthis]

I'm not at all informed on the paleontology, but I've seen multiple datings that suggest most or even all regularly recognized modern bird orders crossed the dino extinction event.

I even chose the date originally specifically because it was the oldest date I could find at the time (I started this sometime in 2010) that would still retain a singular Passeriformes (although I've recently seen a paper suggesting that the split of Acanthasittidae may even be older than that).

OK, for the thought exercise purposes, here goes. Only extant families & orders included.

Struthioniformes - 1 family
Rheiformes - 15 families
Anhimiformes - 1 family
Anseranatiformes - 1 family
Anseriformes - 7 families
Megapodiiformes - 6 families
Galliformes - 20 families
Gruiformes - 16 families
Charadriiformes - 18 families
Scolopaciformes - 39 families
Musophagiformes - 3 families
Otidiformes - 8 families
Mesitornithiformes - 2 families
Pterocliformes - 5 families
Columbiformes - 29 families
Cuculiformes - 22 families
Caprimulgiformes - 8 families
Steatornithiformes - 1 family
Nyctibiiformes - 4 families
Podargiformes - 5 families
Apodiformes - 19 families
Phoenicopteriformes - 4 families
Opisthocomiformes - 1 family
Phaethonhtiformes - 1 family
Eurypygiformes - 2 families
Gaviiformes - 2 families
Ciconiiformes - 3 families
Suliformes - 4 families
Threskiornithiformes - 9 families
Pelecaniformes - 18 families
Sphenisciformes - 1 family
Procellariiformes - 17 families
Coliiformes - 4 families
Leptosomiformes - 1 family
Trogoniformes - 6 families
Bucerotiformes - 9 families
Coraciiformes - 11 families
Piciformes - 18 families
Strigiformes - 16 families
Cathartiformes - 1 family
Accipitriformes - 20 families
Cariamiformes - 1 family
Falconiformes - 3 families
Psittaciformes - 22 families
Passeriformes - 194 families

I think the main point of interest in the exercise is how 'off' current taxonomy might be considered in the gross discrepancies of how it treats different groups. Even within Passeriformes, there are areas where I have multiple family-level splits (antbirds come out to 7 families under this system) while others groups are much lumped by the same standard (only 4 families of 9-primaried oscines). To what extent the heavy splitting of non-passerine groups here is an artifact of applying a single standard to longer-lived groups I do not know, but it does at least lead me to ponder if a system that properly adjusted for evolutionary rate would still lead to the large families we currently see or instead suggest making at least some of these splits.

Very interesting!🤔
I must ponder over these as soon as I have some time

 

[THE_FERN]

You lump together taxa very dissimilar biologically and ecologically, and split much more similar taxa into two or more very similar groups. Basically, you have shown that time-calibrated, monophyletic groups are an impractical method for classification of birds.

It fits well the current understanding that evolution proceeds at different speeds depending from the strength of the natural selection. In this context, interesting is that birds living in the ecologically stable rainforest environment produce many deep genetic divergences similar ecologically (forest partridges, forest rails, forest falcons, nightjar relatives), while birds of unstable open habitats and waterbodies produced relatively many dissimilar but relatively closely related groups.

I feel you could create a classification of birds which is more practically usable and more likely to be adopted if you changed the approach. Define groups which are monophyletic in a narrow sense of having a common ancestor and including some descendants, and defined by morphological and ecological similarity. This method is objective and replicable. It requires a list of morphological characters and finding groups most similar within each other but most dissimilar between each other. It can include extinct species which genetic information is not accessible. It will also produce groups which are practically useful, not ones linking species which have visually, biologically and ecologically little in common (likes of the honeyguide-woodpecker group, flamingo-grebe group etc.).

The nub of this is "what is the point of the classification"?

Basically, from a scientific point of view a time-based approach (perhaps where time is measured as generations) is the only way you'll get "equal" taxa. And here "equal" means equal opportunity to diverge---whether they have done or not. This surely provides a stronger basis for scientific comparisons.

If, however, the object is a "functional" classification, then equivalence of taxa isn't important (not that it's even possible to measure)---perhaps the objective here is to produce readily recognisable groups.

Current higher taxonomy "sort-of" follows this second approach making it a poor framework for science. There is, of course, no reason why you can't have more than one taxonomy for different purposes...

 

[THE_FERN]

On the contrary, I would argue this shows that using objective criteria produces different results than using subjective criteria such as 'similarity' (to the human eye). Which criteria are preferred is another question entirely.

The second, the similarity thingy, isn't in any sense objective or repeatable. It could be made more so, but only with respect to a limited set of dimensions of difference---not the whole set

 

[Lophostrix]

I love "Silver Plover". Nice concept which fits well with impressions the summer plumage lends.
Golden lapwing is just weird🙃

I also like Silver Plover, maybe we should adopt this as an universal name. It would be a good way to retire the American/European name differences.

 

[njlarsen]

I also like Silver Plover, maybe we should adopt this as an universal name. It would be a good way to retire the American/European name differences.

except that it only applies to 25% of the species ...
Niels

 

[Lophostrix]

except that it only applies to 25% of the species ...
Niels

I should clarify I was talking about a good way to retire this particular difference