Whistling/Humming as Evidence

I uploaded my recent observation to see how people felt about this.


There are numerous threads on the forum talking about drawings as evidence
and so far that seems to be upheld by iNat. I couldn’t find anyone asking about similar recreations of a sound. So here is my best recreation of a bird sound that I heard but couldn’t get a recording. I personally think this is just as “verifiable” as drawings. I certainly understand peoples’ hesitation to accept drawings for RG on iNat, but currently they are accepted. I am curious if this kind of observation could follow suit


Some will disagree, but seems fine to me (especially since imo a call reconstructed from memory will probably be more accurate than a drawing reconstructed from memory)


Agreed. I made sure to do the recording as soon as possible after hearing it. Recreated observations will of course be limited to the skill of the “artist”. But this is relatively accurate to what I heard

1 Like

If drawing’s fine I don’t see why this wouldn’t be. Same idea, different medium.


Some people might be tempted to identify the observation as Homo sapiens, so avoid misunderstandings, I would continue to add a description, and a comment after your initial ID, making it very clear what the sound is intended to represent.


This is more prone to misidentifications than a drawing where you can at least show what you saw schematically, you can’t whistle the way you’re able to separate coal tit from goldcrest, so unless the call is so distinctive you can’t misid it, that’s not the best idea to try to imitate birds.


All those points make sense, but realistically the same principle applies to drawings and even pictures. If it’s a bird (or anything else) that might need sound or behavior or a minute detail, the drawing or picture won’t provide enough evidence.

Maybe mine will or won’t be ID’d. But to me the practice seems fine.


Yup, and many drawings stay at high level, with audio there’ll be more situations where people can hear something different from what was heard, with normal bird recording you at least have its voice, can check the sonogram, with human voice I’m afraid there’ll be a lot of misids, esp if observer doesn’t know the bird. I see both types of recording more suitable if a) observer 100% knows what it is and this knowledge can be proven b) object is sitting for a long time, enough to make a good sketch of all details needed for an id.
Anyway, upload those and we will see if they’re idable, just be cautious.


Interesting idea, but I have my doubts. I think my main reservation is that with a drawing, it is immediately clear it is a drawing, and identifiers can move on if they wish. With a recording, it will be saved in the database alongside audios of the species itself, and it is harder to know that it does not contain direct evidence of the species without actually listening to it. I wouldn’t be super keen on seeing more observations like this, and it would probably make me, as an identifier, even more reluctant to listen to audios if they became more common.


As someone who enjoys identifying bird audio on iNat, this idea meets our needs for creativity and fun!

From a data quality standpoint, probably the same rules for drawing: Was made where the observation occurred and right away after hearing.

We might also not want to encounter this widely.

We speculate that this is unlikely to become a widespread phenomenon. And we had fun trying to figure out what your bird could be.


I wonder how many of you could tell the difference between me imitating a Red-tailed Hawk and me imitating a Steller’s Jay imitating a Red-tailed Hawk?


same could be said of actual audio of bird calls tbh. Unless the bird was actually seen while calling, then theoretically any bird call uploaded from the east coast of Australia could actually be a superb lyrebird mimicking the apparent species


It seems like it should be fine, but echoing those data quality concerns, one wonders if perhaps there needs to be a change in iNat policies to make those types of observations (drawings and repeated… audio… things?) have some kind of flag/tag so that they can be excluded in research that requires stricter quality standards. That would be a decent compromise between allowing it while not allowing it to interfere with data quality, but it’s dependent on that change being implemented…

this strikes me as such an odd concern. over in eBird, people generally aren’t asking for or providing audio recordings or photos, and yet that data gets used for research all the time.


I like this idea, and have a perverse urge to see how far it can be pushed.

If drawings and imitations of sounds qualify, why not textual descriptions of what traits were observed? I’m terrible at drawing, and would struggle to make a drawing that incorporates the traits I observed, but I can tell you exactly what characteristics I observed that lead to my ID. For example, today in my yard in Santa Rosa, California, I observed, under a small log I was moving, several extremely long slender salamanders. Each was dark brown with the slightly lighter brown stripe along the back, textured such that it looked almost like they had some scales on their backs. They had very moist skin, roughly 3mm wide at the chest, about 4cm long, cylindrical, with truly tiny legs and feet, a wide, short rounded snout, and round black eyes. Some of them curled up into tight spirals when disturbed. Many naturalists in my area could ID this to species without hesitation, and several could draw it accurately based on this description. Why then, should a drawing be evidence but an image of text not?
EDIT: I have submitted this observation as a test case: https://www.inaturalist.org/observations/146777425

Give me evidence of an organism. If I can’t see anything other than a picture or a noise from the critter as evidence I don’t entirely trust it. It strikes me as unreliable, therefore undesirable. It’s lovely that you did it, but I don’t necessarily believe you.

I want a photo or a recording of the critter.

Personally, I do not think I would have any use for observations based on either drawings or attempts to recreate sounds, nor would I provide IDs for them. If these kinds of observations were common enough to be “in the way”, I would want a way to easily separate them out from the pile. Otherwise, they just fall into the general category of people occasionally doing things that I think are a little odd. :-)

1 Like

I’m going to post my whistled evidence of the Loch Ness Monster meeting a Yeti to eat a Chupacabra. Hand-drawn evidence will follow.

I’d argue drawings of animals are subject to similar faults. If you saw only a blur and decided to draw the bird you believe you saw, there’s the same risk of bias as there is the risk of generalization with recreated sound observations. Most people drawing the animal or plant they saw would have a general idea of what lives near them, and that can influence the decisions made when creating such an art piece.

And that is why you need to draw what you saw, not what you think you saw.

1 Like