Subspecies groups and future splits (1 Viewer)

[Mysticete]

Have you also tried mapping in the BirdLife Taxonomy?

When I did this I found that there were several splits that were not covered by Clements groups. I will have a check through my mapping (although it now needs updating to BirdLife 8) and then through this thread to suggest any other potential splits.

Actually from my mapping, I have created my own set of groups (for each taxonomy) which is the minimum level or granularity required to enable exchange of data at species level between IOC, Clements and BirdLife… so not groups for Common Merganser (American) and (Eurasian), but under Clements the retention of Green-winged Teal (Green-winged) and (Eurasian), plus as I say some newly created groups to fit some splits in other lists.

It has been fascinating (but a bit tedious) to see that generally convergence is occurring and that the number of these (minimum requirement for species mapping) groups, is reducing year by year.

Yes...I have the "All the birds of the World" book from lynx and have incorporated the subspecies groups from there as well, when they recognize different ones from Clements (or splits not representing bird groups not recognized in Clements). I also try to go through Opus, Birds of the World, and Wikipedia to look for other potential splits/groups.

Feel free if you have "groups" for the already covered species that are not in my file to share them here!

 

[Jon.Bryant]

Yes...I have the "All the birds of the World"

Have you tried downloading the BirdLife Taxonomic list, which has updates post AtBotW? See here

BirdLife list link

I recall an example of a missing group is Cercococcyx mechowi and lemaireae Dusky and Whistling Long-tailed Cuckoo. I think (from memory) that IOC and Clements treat Dusky as monotypic and do no recognise Whistling, whereas BirdLife recognise two species based on vocalisations only (and possibly differences in tail length). I think the two species were recognised post publication of the book.

As I say I will do some leg work and update, so I am not talking out of my hat.

 

[Mysticete]

Have you tried downloading the BirdLife Taxonomic list, which has updates post AtBotW? See here

BirdLife list link

I recall an example of a missing group is Cercococcyx mechowi and lemaireae Dusky and Whistling Long-tailed Cuckoo. I think (from memory) that IOC and Clements treat Dusky as monotypic and do no recognise Whistling, whereas BirdLife recognise two species based on vocalisations only (and possibly differences in tail length). I think the two species were recognised post publication of the book.

As I say I will do some leg work and update, so I am not talking out of my hat.

Maybe I am missing something, but does the file in that link contain subspecies groups (rather than just subspecies)? I scrolled through it and didn't notice anything. Hence why I am using Birds of the World

 

[Paul Clapham]

You're correct; the Birds of the World spreadsheet contains a (presumably complete) list of subspecies. What I see in the book ATBOTW is drawings of a short list of subspecies for each species. However if you look at e.g. Synallaxis rutilans it says "7 sspp in 2 groups" and there are 3 drawings. Sometimes each of the drawings in such a case has a small red number next to it to indicate the group number, although that isn't the case for e.g. Melanochlora sultanea. The groups themselves are not named and their contents are unspecified. Anyway I'm sure you'll be able to figure it out.

 

[Jon.Bryant]

Maybe I am missing something, but does the file in that link contain subspecies groups

No unfortunately not… it is a shame as HBW Alive did have subspecies groups, and I don’t know why BirdLife didn’t use them (the also discarded all the range info in HBW Alive, which is a real shame … I extracted this data by writing a webscraper (prior to the death of HBW Alive) and created a version of their taxonomy with range info, but they haven’t adopted it. There reply was ‘it would unfortunately mean more work to keep updated’.

I actually keep meaning to contact IOC and BirdLife with my ‘must have for mapping’ groups, to see if they would consider adding for user benefit/data exchange between lists - I suspect the WAGC could one day make this redundant (as well as my mapping!).

Anyway, I have mapped every species and subspecies in IOC (no groups), Clements and BirdLife (also no groups). I say map… it is an Excel spreadsheet with a single row for each taxa (species or subspecies), and I have recreated the taxonomies and mapping sheet in a SQLite database. This is a bit different from the IOC comparison spreadsheet, which has a separate row for each species and then indicates different spellings and who recognises the species. With my sheet all mapping data can be read across a single row…

I have added in the Clements groups to the sheet, which is why I know that Clements groups are not entirely sufficient to map all splits and lumps.

Unfortunately by moving to SQLite, I now have to write database update programmes to capture the latest updates (Clements 23, BirdLife 8 and IOC 14.1), so my database is currently not entirely up to date. This means that if I extract the relevant info for you, it may have already been ‘fixed’, by one of the latest updates.

ATBOTW is good and shows where main taxonomies vary in opinion, but I think it was compiled in 2018, and BirdLife (plus IOC and Clements) have made annual changes since (BirdLife every December). I admit that recently BirdLife changes have not been as wide sweeping as Clements and IOC (.. I have yet to analyse the 2024 changes, though), but ATBOTW is slowly becoming out of date, so it is not really reliable as a primary source.

 

[Jon.Bryant]

BTW when I spotted a minor error in the IOC list (inclusion of a tab character at the end of the range data, which was sending my Python and Pandas programme screwy), I had to explain to IOC how to delete the character from the master Excel sheet… I was really surprised that they ARE just using Excel and don’t have a database back ene populating the spreadsheet.

Excel has proven time and time again to be an error prone solution as it does not allow separation of UI, data access layer and data and does not comply with DRIP (do not repeat yourself) principles.

IOC seem to do a good job of error checking, but I keep finding more important errors in the BirdLife Data (e.g a Bluethroat subspecies that had accidentally be given the genus name (with capitulation), rather than the subspecies name.

The only kind of error I have found so far in Clements is and extra tab space in one line and that for Red Crossbill there is no group for nominate birds (presumably this should be Red Crossbill (Red).

 

[kweetal]

BTW when I spotted a minor error in the IOC list (inclusion of a tab character at the end of the range data, which was sending my Python and Pandas programme screwy), I had to explain to IOC how to delete the character from the master Excel sheet… I was really surprised that they ARE just using Excel and don’t have a database back ene populating the spreadsheet.

Excel has proven time and time again to be an error prone solution as it does not allow separation of UI, data access layer and data and does not comply with DRIP (do not repeat yourself) principles.

IOC seem to do a good job of error checking, but I keep finding more important errors in the BirdLife Data (e.g a Bluethroat subspecies that had accidentally be given the genus name (with capitulation), rather than the subspecies name.

The only kind of error I have found so far in Clements is and extra tab space in one line and that for Red Crossbill there is no group for nominate birds (presumably this should be Red Crossbill (Red).

The huge advantage of using a spreadsheet (over a database) is that you don't need database people. Do you realize what they cost?

 

[Jon.Bryant]

presumably complete) list of subspecies.

Yes correct. One of the great things about the BirdLife list is that there is an accompanying taxanomic notes spreadsheet, which explains exact why they do and do not recognise certain species and subdpecies.

 

[Jon.Bryant]

Do you realize what they cost?

Perhaps an issue for IOC, but is money a real issue for Cornell (and perhaps Birdlife).

Also we are not talking about a large relational database with web front end, just something simple to act as a master and export to the familiar Excel sheet … even MS Access would suffice, which is easy for even a moderately interested self taught person to master.

 

[Mysticete]

ATBOTW is good and shows where main taxonomies vary in opinion, but I think it was compiled in 2018, and BirdLife (plus IOC and Clements) have made annual changes since (BirdLife every December). I admit that recently BirdLife changes have not been as wide sweeping as Clements and IOC (.. I have yet to analyse the 2024 changes, though), but ATBOTW is slowly becoming out of date, so it is not really reliable as a primary source.

For the scope of this sort of project, I think ATBOTW isn't too out of date. The the major differences between it and more recent versions is the addition of some splits, which are easy to accommodate since they often result in a subspecies group graduating to species. I also try to use additional sources beyond either Clements or ATBOTW, so hopefully I am capturing all of the known future splits, although as always I more than welcome people volunteering any additional ones I might have missed!

 

[kweetal]

Perhaps an issue for IOC, but is money a real issue for Cornell (and perhaps Birdlife).

Also we are not talking about a large relational database with web front end, just something simple to act as a master and export to the familiar Excel sheet … even MS Access would suffice, which is easy for even a moderately interested self taught person to master.

I don't know, the data is really tiny and more importantly, it hardly changes, and then mostly after much deliberation .

It's going to be very hard to capture taxonomic expertise into constraints, so what advantage would a database bring?

 

[Paul Clapham]

IOC seem to do a good job of error checking, but I keep finding more important errors in the BirdLife Data (e.g a Bluethroat subspecies that had accidentally be given the genus name (with capitulation), rather than the subspecies name.

I found one entry on a BirdLife spreadsheet from a few years ago which had a non-breaking space in the middle of a name. Took me a long time to figure that one out.

 

[Jon.Bryant]

I don't know, the data is really tiny and more importantly, it hardly changes,

If you include all the subspecies, then roughly 30.000 entries, so I think really tiny is not quite apt. IOC create a spreadsheet with a lot of white space and no repetimos, but Clements and BirdLife repeat Order, Family and Latin name throughout their spreadsheets.

I am currently working through changes to Clements, made in the last update, and if you go down to subspecies level there are a little short of 800 changes. to Latin names alone (including deletions and additions). So again I would argue that ‘hardly changes’ is not really right. In fact yeses 800 changes are to my database —- depending on the structure of a spreadsheet there will be more changes caused by repitión,., one Genus change not only can effect the species entry, but the same Latin nsme change then needs to be copied down to each and every subspecies row.

I would argue that not only is a database less likely to contain errors (you can run data validation on input), but also quicker to update once created … I.e you make a would once without having to copy and paste to every impacted row! I think in most walks of life it would be the natural solution (and surely Clements and BirdLife with their large online data centres, do actually administer other databases?)

 

[Paul Clapham]

The recent Clements checklists show every sign of being generated from a database. Which is not surprising considering the number of database specialists they must have on staff. The latest checklist did have subspecies Terpsiphone atrocaudata illex belonging to subspecies group Terpsiphone atrocaudata atrocaudata/ilex but that's not the sort of thing that validation would catch easily. (Turns out that illex is correct.)

But the recent BirdLife checklist wasn't. It has the order and family information on every line for a species. That's redundant from the database normalization point of view, sure, but you could do it in the process of creating the spreadsheet from a database. However the family column has both "Typical Owls" and "Typical owls" in it. (Alternate possibility: their database isn't normalized.)

I'm sure if the list producers gave everybody who reads this forum the option to download some kind of database or an Excel spreadsheet, pretty much everybody would take the spreadsheet.

 

[kweetal]

If you include all the subspecies, then roughly 30.000 entries, so I think really tiny is not quite apt. IOC create a spreadsheet with a lot of white space and no repetimos, but Clements and BirdLife repeat Order, Family and Latin name throughout their spreadsheets.

I am currently working through changes to Clements, made in the last update, and if you go down to subspecies level there are a little short of 800 changes. to Latin names alone (including deletions and additions). So again I would argue that ‘hardly changes’ is not really right. In fact yeses 800 changes are to my database —- depending on the structure of a spreadsheet there will be more changes caused by repitión,., one Genus change not only can effect the species entry, but the same Latin nsme change then needs to be copied down to each and every subspecies row.

I would argue that not only is a database less likely to contain errors (you can run data validation on input), but also quicker to update once created … I.e you make a would once without having to copy and paste to every impacted row! I think in most walks of life it would be the natural solution (and surely Clements and BirdLife with their large online data centres, do actually administer other databases?)

The IOC master list has 33,682 rows, and yes, I do think that's tiny (a few MB). Likewise the changes over a six months period are limited; 800 changes is not a serious load (800 writes per second is a serious load).
Database strong points are: (1) handling huge data, (2) concurrent access, (3) enforcing constraints. In the IOC case, I don't see any of these apply.

Anyway - ours not to reason why

 

[THE_FERN]

The IOC master list has 33,682 rows, and yes, I do think that's tiny (a few MB). Likewise the changes over a six months period are limited; 800 changes is not a serious load (800 writes per second is a serious load).
Database strong points are: (1) handling huge data, (2) concurrent access, (3) enforcing constraints. In the IOC case, I don't see any of these apply.

Anyway - ours not to reason why

Those are some but not all of the plus points for a typical database. Certainly, compared with a spreadsheet a proper set manipulation language (SQL) is one.

Actually, the appropriate storage for taxonomic data is a graph database.

 

[kweetal]

Those are some but not all of the plus points for a typical database. Certainly, compared with a spreadsheet a proper set manipulation language (SQL) is one.

Actually, the appropriate storage for taxonomic data is a graph database.

Perhaps we should have a separate technical thread about how one can use databases, for names.
Someone should try Apache AGE, the postgres graph extension, and see what it can do with names.

 

[Mysticete]

Perhaps we should have a separate technical thread about how one can use databases, for names.
Someone should try Apache AGE, the postgres graph extension, and see what it can do with names.

Indeed....this has wondered pretty far from the topic...

 

[Björn Bergenholtz]

Just a quick observation, and a short comment (alt. a minor remark) ...

Re. your post #137, and the "Final batch of Nocturnal Strisores":

[...]

Mountain Owlet-nightjar
Albert's Owlet-nightjar - Aegotheles (albertisi) albertisi
Archbold's Owlet-nightjar - Aegotheles (albertisi) archboldi

If eponyms are to be used , note that dedicatee of the former wasn't an 'Albert' at all, but an Italian Count (Conte), by the name (and title) Luigi Maria (Conte) d’Albertis, thus (at least in my mind) the first one ought to be (either):

Albertis's Owlet-nightjar Aegotheles albertisi P.L. Sclater 1874, or (Count) d’Albertis' Owlet-nightjar, alt. d’Albertis's Owlet-nightjar (or likewise, whatever version you prefer). See Jobling's Key (here):

albertisi / albertisii
Luigi Maria Conte d’Albertis (1841-1901) Italian botanist, zoologist, ethnologist in the East Indies and New Guinea 1871-1878 (Aegotheles, Drepanornis, syn. Goura scheepmakeri, Gymnophaps).

That is if today's Common name is connected to [Philip Lutley] Sclater's Scientific name of course ... (it's also been called Arfak - alt. Mountain ditto).

As far as I know there hasn't been any (true) 'Albert' associated with this Bird.

But if so, please enlighten me (and us all).

Either way: Good luck with your many lists/names.

Cheers

Björn

 

[Mysticete]

Finally getting to the Hummingbirds (so so many hummingbirds...). It will take me some time to get through the potential splits/subspecies groups for these

First up are the hermits:

Pale-tailed Barbthroat
Pale-tailed Barbthroat - Threnetes (leucurus) leucurus
Bronze-tailed Barbthroat - Threnetes (leucurus) loehkeni

Tawny-bellied Hermit
Train-bearing Hermit - Phaethornis (syrmatophorus) syrmatophorus
Colombian Hermit - Phaethornis (syrmatophorus) columbianus

Long-billed Hermit
Long-billed Hermit - Phaethornis (longirostris) longirostris
Baron's Hermit - Phaethornis (longirostris) baroni

Mexican Hermit
Mexican Hermit - Phaethornis (mexicanus) mexicanus
Jalisco Hermit - Phaethornis (mexicanus) griseoventer

Great-billed Hermit
Great-billed Hermit - Phaethornis (malaris) malaris
Amazonian Hermit - Phaethornis (malaris) moorei
Margaretta's Hermit - Phaethornis (malaris) margarettae

Straight-billed Hermit
Straight-billed Hermit - Phaethornis (bourcieri) bourcieri
Ash-bellied Hermit - Phaethornis (bourcieri) major

Pale-bellied Hermit
Pallid Hermit - Phaethornis (anthophilus) anthophilus
Dark-crowned Hermit - Phaethornis (anthophilus) fuscicapillus

Streak-throated Hermit
Boucard's Hermit - Phaethornis (rupurumii) rupurumii
Santarem Hermit - Phaethornis (rupurumii) amazonicus

Cinnamon-throated Hermit
Cinnamon-throated Hermit - Phaethornis (nattereri) nattereri
Maranhão Hermit - Phaethornis (nattereri) maranhaoensis

Reddish Hermit
Red-vented Hermit - Phaethornis (ruber) ruber
Bishop Hermit - Phaethornis (ruber) episcopus
Black-banded Hermit - Phaethornis (ruber) nigricinctus

Gray-chinned Hermit
Gray-chinned Hermit - Phaethornis (griseogularis) griseogularis
Porculla Hermit - Phaethornis (griseogularis) porcullae

Stripe-throated Hermit
Stripe-throated Hermit - Phaethornis (striigularis) striigularis
Dusky Hermit - Phaethornis (striigularis) saturatus