Identifying birds to subspecies without field marks

There is a frustrating trend among birders using iNaturalist toward identifying birds to subspecies even when they have no field recognizable field marks. Many of us have had this discussion off-forum as it relates to Song Sparrows, Fox Sparrows and juncos, but it seems to be spilling over into other species where the subspecific designations are based on museum tray work and have no consistent or reliable characters that can be used in the field.

Birds have wings and cannot be reliably be parsed to subspecies by range alone, especially in migration. One would think that birders, with their obsession over vagrants and rarities, would get that. I have no issue with ID based on recognizable forms, but range is a slippery slope and most of the folks making their claims are not even citing a reference for their opinionā€¦

6 Likes

Iā€™ve noticed this too especially with birds in the neotropics and Africa.

Based on my experience, youā€™re unlikely to ā€˜winā€™ the argument with them, or convince them to stop, as they will insist they are being more scientifically rigourous or pure than you. It comes down to if you feel it is worth your time to enter a dissenting ID at just the species level.

EDIT - I should point out that there is a school of thought, although I do not think it is formally documented in the Terms if Use or Help pages that if you can not prove a userā€™s ID is incorrect, it is inappropriate to enter a dissenting ID.

2 Likes

In herps there is also a large segment of the identifier population which IDs to subspecies based only on range. I too find this frustrating as (at least in herps) subspecies are often not very scientifically sound (many are overturned quite frequently). Additionally, if the identification to subspecies is based solely on range, it does not add any information to the observation. The locality is already included, and anyone who is interested in subspecies could make that determination for themselves!

And as @mikepatterson pointed out, organisms (especially birds, but other things as well) do move or are transported long distances. IDing to subspecies based only on location and not on characters means that some interesting and valuable instances of long distance migrants will be missed. IDing to subspecies based solely on range can also make taxon swaps and updating more difficult when subspecies end up not corresponding to a newly described species, so there are potential costs to this.

I think this is a representation of a general trend of many IDers to be biased towards offering the more/most specific identification, even when it may not be warranted/supported by the available evidence. As @cmcheatle noted, itā€™s unlikely that anyone will ā€œwinā€ this argument due to folks feeling that the highest level of precision is always the best.

I do not, however, think that itā€™s inappropriate to enter dissenting IDā€™s (though I have heard some argue otherwise). Itā€™s probably good form to explain why you do so. However, in an egalitarian system like iNatā€™s IDing, if folks who think an ID is unwarrantedly specific donā€™t dissent, then that voice isnā€™t heard within the system, and an inherent bias will result.

1 Like

Part of the thought is that the best to do is ignore it. For example there is a fungi genus found in Ontario, for which there are 2 members in the province. They can only be identified by microscopic spore review or dna sequencing. Yet inevitably they get entered under 1 of the 2 species (it has a common name, the other does not, it has a bunch of records, which are a self-fulfilling circle).

The ā€˜issueā€™ is that if a species is entered:

  • I canā€™t prove the user is wrong, even if it is a guess that is correct
  • however, the user canā€™t prove they are right.

So who gets the benefit of the doubt ?

1 Like

https://forum.inaturalist.org/t/notifications-of-subspecies/2451/5

there have been other posts about this too, one about a snake i think which is identified to subspecies based on which side of a river it is on. I agree that it shouldnā€™t be done at least at subspecies level, at best itā€™s pointless because people can see where the observation is anyway, at worst it creates false data because you might be missing a surprising population of the other subspecies.

In general IDā€™s based on range are a slippery slope. Not Ok for subspecies, but Ok for species as an example.

I remember an interaction I had on the site, where I entered a record of an Eastern Wood-Pewee (thatā€™s a flycatcher bird for non birders). A user came and asked did I hear it singing, I replied no, so they bumped it back to genus saying Eastern Wood-Pewee and Western Wood-Pewee are indistinguishable other than by song, so therefore my ID was invalid, and lacked supporting evidence for anything but a genus ID.

Even pointing out to them that Western Wood-Pewee has been documented in the province I live once (technically 3 times but the other 2 were a good 2,000 kilometers away), they were insistent that anything other than a genus ID was not valid as it was not possible to prove I had not found the 2nd.

The point being, as with virtually everything in terms of ā€˜evidenceā€™, IDā€™s etc it comes down to judgement.

5 Likes

Kelp Gull Larus dominicanus. Southern Africa: Larus dominicanus vetula, Australia Larus dominicanus dominicanus. Are they the same?

I almost never turn off community id but I will when people get absurd that way.

3 Likes

IDs are made ā€œon rangeā€ all the time, and in many cases this is not an issue because a lot of them donā€™t overlap. If taxonomy changes and rankings are obsolete, the subspecies get merged into the species anyway.

For instance: ā€œCaliforniaā€ pipevine swallowtail is identified solely on range. So are several other plants, insects, so on.

The question then becomes how you draw the line between being careful, and being confident, when it comes to subspecies where visual fieldmarks are no good.

2 Likes

I think a lot of it really comes to a case-by-case basis. In a few cases, there may be ranges where only one subspecies is known, and thereā€™s as good evidence of the absence of the others as is possible (wasp example: Polistes dorsalis californicus is essentially the only subspecies found in at least 95% of California). In other cases, there may be known subspecies that not only have no overlap in the US but are distinctly separated (wasp example: Polistes comanchus comanchus from Texas and Polistes comanchus navajoe from Arizona). Then there are cases where subspecies blatantly overlap, and there may even be intergrading between subspecies (wasp example: Polistes dorsalis dorsalis and Polistes dorsalis neotropicus overlap and intergrade in Texas).

Having that sort of familiarity with trends outside of the particular ID is probably a key point, whether weā€™re dealing with wasps or birds. One additional point is the ratio of these species within zones of overlap: are we talking 50/50 or 95/5? The latter would probably be safer to just go ahead and ID whereas the former would be more questionable. Then there are cases of 10/10/10/10/10/10/10/10/10/10 (10/90 for your ID) where your likelihood of being wrong would outweigh your likelihood of being right.

Disagreeing IDs (explicit disagreement) really arenā€™t understood uniformly through the site, and they seem to be used a bit differently depending on the taxonomic group. My main thought would be asking what the likelihood is that itā€™s actually the other (if itā€™s literally a toss-up at 50/50, it probably shouldnā€™t be considered a valid ID). If we consider the concept of falsifiability, one of the requirements is that it must actually be testable (so if traits required for ID were neither recorded nor observed, it wouldnā€™t really be a falsifiable statement). With birds, thereā€™s probably a bit more in terms of falsifiability (and fewer options, maybe even just 2 options), though there are tons of insect IDs that simply arenā€™t falsifiable from most photos (thereā€™s one cryptic species complex with 15 options thatā€™s literally impossible to ID by photos, much less any existing subspecies).

  • With the pewee situation, you seem to demonstrate the point of probability, provided those other records had gone through enough to pass the falsifiability test. That being the case, it would be arguable that the population had passed diagnosis.
1 Like

Itā€™s more of a logic thing, really. If they say it is A. b ssp c, but you think it canā€™t be IDd beyond species, then you ID as A. b without explicit disagreement. If you can see a reason WHY it canā€™t be ssp c, then you ID as A. b with explicit disagreement. If you can acknowledge that there is the POSSIBILITY that it could be A. b ssp c then you should NOT be explicitly disagreeing.

If you think the other identifier is wrong, then instead of ā€œpushingā€ the ID to what you think it should be, you should start a dialog with them, help them see your perspective and encourage them to reconsider their ID.

It is called a ā€œCommunity IDā€, which is come to by factoring in everyones IDs. If you start ā€œLyingā€ with your own IDs just to make CID become what you personally think it should be, then consider how you would react if they in turn enlisted 5 dummy accounts to ā€œpush backā€ and make the ID what they think it should be. Just make your ID as you see it, and let the Community ID fall as it may.

1 Like

The problem I have is that while I may ID as A. b without explicit disagreement, some less judicious identifiers will go ahead and agree to A. b ssp c even though they have no support for that level of ID. They simply agree to anything that seems remotely plausible. (Iā€™ve even seen users who agree to species on photo-less obs that they clearly have no way of verifying!) Now the obs is RG at A. b. ssp c even though the submitted evidence does not warrant that level of ID.

I have mixed feelings on this. Going back to the 2 fungi example in my earlier post in the thread. A user enters a record at the species level, and has not done the DNA or microscopy. There are 3 possible outcomes to this record
1 - Another user comes along and agrees to the species id. You get a research grade record at the species level that might be right, but equally could be wrong with no way to prove either
2 - Every user ignores it and it stays at needs id in perpetuity
3 - Someone enters an id at the genus level notes why and recommends using the no it cant be improved and you get a research grade genus record

To my mind the preferred hierarchy of outcomes from above is 3,2,1 but I recognize others may feel differently.

EDIT - I think Jane and I were writing at the same time.

4 Likes

You said it better though!

Well, since it canā€™t achieve research grade anyway it seems pretty harmless, maybe just an ā€˜i agree that species is thereā€™ or even ā€˜i also saw that species in that locationā€™.

There is one iNaturalist user who blocked me from commenting on any of his IDā€™s because I habitually restricted my input to species level with explanations for my reasoning. It makes me wonder how @fogartyf has managed to do the clean up heā€™s trying to do with juncos without getting yelled at :slightly_smiling_face:

1 Like

probably they got yelled at :(

There is at least 1 reputable use of this, for example the project admins of the provincial rare species tracking project where I live will do this with users they know are trustworthy as a marker that they consider the record appropriate for inclusion in the project and tracking dataset.