What is evidence?

If the subject organism (a virus in this case) is causing some diagnostic behavior, change in appearance, etc. in another organism and that is exhibited in the photo, then I think it’s okay. However, I think the evidence would be best classified as “sign” and not the actual organism.

I’ve seen photos of dead prairie dogs that reportedly died due to sylvatic plague from Yersinia pestis. Possibly that might work as an iNat record of Yersinia, although the evidence would have to be supported by lab testing of the carcass to confirm cause of death. But in reality there’s no way for an independent reviewer to see something in that photo that says Yersinia.

I don’t think photographing a multicellular organism that happens to carry some bacterium or virus (but doesn’t show evidence of carrying it in the photo) would count as an observation of that microorganism.

3 Likes

Hi
Nice to see this discussion, esp coming in light of my own upcoming project to “encourage” people to post old observations.

So apart from detailed field notes / sketches with descriptions etc what do people think about

  1. ebird checklists used in the reference to a bird observation
  2. Hand written checklists made on pre ebird trips (again for a single organism)

As an ebird reviewer I know that many “species” are automatically accepted due to filters setup by experts, on the other hand many species are not accepted / not confirmed for the same reason. While some don’t even appear in the public domain (marked not for public).

But ebird is largely based on honesty tempered with “reviewers” cross checking both species level regional filters, and a user trust system based either on history and / or personal or professional interactions with the observer.

Personally, I think I would not use an ebird checklist reference to make an bird observation in inaturalist but do think it may be a feasible idea.

Very good point and a very relevant one in the world of birding

2 Likes

to add more reason to why i would, personally, not use ebird to make an observation on inaturalist. Because

  1. ebird records are adequate for me - to know that the bird has been recorded for posterity and it need not be duplicated on inaturalist just for the sake of species building etc
  2. I have a camera and sound recorder so would ideally want to share something interesting

To re-iterate I would not be averse to “helping” others identify a species by ebird checklist and description if they would like to do so.

Based on this, it is probably a reasonable assumption that there is some additional number of eBird checklist taxa that were also wrong, but happened to pass the expert filters and so were not caught. That seems like an additional good reason not to use an eBird checklist as the sole evidence for an iNat observation.

That is true,
some of them do not meet the standards set and are marked “not for public” and this data does not make it into the public domain.

Usually this is people who are either new to ebirding and start putting on their entire life list on one checklist or the usual group of people for who showing off what they have seen, even by fakery, is the norm.

My point was not the entire checklist but certain species that are “rare / unusual” etc - these normally get flagged by the filters and mostly are then vetted for reliability and accuracy.

While ebird is a massive the platform by and large the reviewers are from regional levels who in turn know many people though various interactions, or can see their history of birding .

Of course the ebird checklists are not based on “documentary” based evidence - it would be fairly impossible for ebird to function if that was the case. In fact birding would collapse if digital evidence for each bird was sought. (not to mention a lot of bird research as well).

**
Some grammar and spelling edits made

1 Like

Nothing would collapse really, having a recorder with you working all the tie would be enough to get all the sounds you heard and camera could capture everything else, silent but seen. I see tons of checklists with very doubtful birds seen, but there’s no photo, so it can’t be proved or disproved, or there is a photo and it’s of another species or it’s an american species seen in the middle of Siberia with no photo and description “common”, so it’s pretty safe to assume big part of checklists have wrong data and shouldn’t be used at all. And we see sae on iNat, user ids are far from being right all the time.

1 Like

Am not going to belabour the point, or start pointing out flaws in either ebird or inaturalist. Neither platforms are perfect and both are heavily influenced by its users.

And am, also, not going to “insist” that ebird records be acceptable on inaturalist.

However do seem to sense a resentment for either ebird or for some non good ebird users that inatters may have had less than nice interactions with.

To also say I endure my fair share of less than nice ebird users too :-)

I am helping train nature guides and nature educators – and often also request them to “triangulate” thei observations by linking it with something some one with a better device may have recorded at the same time / location/

1 Like

Even with a camera and recorder constantly running, there are going to be plenty of birds seen well enough to ID but not well enough to provide documentary evidence. Most field research with birds doesn’t actually require such evidence, in fact – point counts and transect counts usually assume that the people doing the data collection are both honest and capable of making the IDs. With regards to e-bird and the quality of data, that’s going to largely depend upon how the filters are set up and how the local reviewers work…

1 Like

Scientific research with professionals that can doubt themselves is one thing, but millions of regular birders uploading what they think they saw is another, of course this model works on statistics, but having 2 types of records, with and withot proof is essential in my opinion, and can be implemented on eBird too. iNat also relies on statistic approach, but at least you can verify the data, without having a trust in observer’s word on it. Also all kinds of counts are very old and somewhat traditional, so of course they don’t ask for photos if 30 years ago they didn’t.

1 Like

Yes, these are different situations. However, most of the problems you described don’t really seem to occur here in the US, from what I’ve seen, and that’s where the filters and reviewers come in. (I know that e-bird initially had trouble finding people in parts of Europe willing to serve as reviewers – I’m not sure, but I think it’s a volunteer position, and it’s a pretty significant time investment…)
In general, any data set is going to be subject to limitations based on how the data is collected, and any researcher that wants to use a citizen-based dataset needs to spend some time learning about how that set is produced and what sort of limitations that’s going to place on the analysis. (Incidentally, I agree that birds that appear to be out of range shouldn’t be included without, at the very least, an excellent description. In some of the historical stuff, a reference to a published description would make sense…)
Also, one note about e-bird – some records are kept out of the public-facing database because of the potential for disturbance of the birds, similar to the way that iNat automatically obscures coordinates for some species.

4 Likes

It is a volunteer position,

Regarding old records - now there is a concerted effort to digitize birding records. In India the records of Dr Salim Ali’s, the father of Indian birding are being uploaded. With transcribing of the detail field notes kepts. The notes and the bird descriptions are fascinating.


edit to correct grammar, sentence structure

1 Like

What is wrong with a good description?

I cannot understand why in iNat a good text is not enough to start a discussion about an ID, and the observation is immediately Casual!

All birding culture was based on a good description, maybe because it comes from a time when photos and recordings were not easily available. Still today it is a culture of trust and honesty that accepts written descriptions.

A good description that allows to ID a bird is even considered a thing that people should learn to do, like a form of art, but that the new generations are losing. When a birder sees a rare bird that as to be proved, like to a rarities committee, a good description is enough and needed.

eBird, even today, if you observe a rare bird, to be accepted by the reviewers, you must have a photo/recording of it, or if you do not have it, a good description is enough. See this example of a Black rail (rare and sensitive), observed by me, that was accepted by the eBird reviewers. I have a sketch that by itself is not enough to reach an ID. What was important was the habitat, bird, and the observation details.

I believe that iNat, not accepting descriptions, is saying to the observers that do not trust them, a media is needed as form of proof, and that the ability to write a good descriptive text is far less important than photos and recordings. It simply is not.

1 Like

iNat and eBird are completely different platforms and don’t have to follow each other’s rules, but nothing stops you from uploading both sketch and discription to iNat and have them “accepted” by another user, who said you can’t add descriptions when description is available for any observation? Description alone without sketch? You can get yourself a casual observation without picture.

1 Like

Use the sketch for iNat. With the description, so birders can give informed agreement.

I was on the Records Committee of the Iowa Ornithologists Union for several years, evaluating written records. Very frustrating. We often had to reject written records that just weren’t all that descriptive, though I think the problem was with the writer, not the observation per se. We generally accepted written descriptions by a guy who was so good he could always write a great description – but did he always see the weird out of range bird, or just see a glimpse and imagine the rest? I like iNaturalist using photos and sound recordings (and rarely drawings) because the identifications can (in theory) be verified by a person who had nothing to do with the observation itself. Let eBird and other groups deal with written descriptions.

7 Likes

Okay so I have been thinking a lot recently as to what exactly counts as evidence, but in a slightly different sense to the way that it is being discussed here. No point in making a different topic because it is similar enough though. In my mind there are a lot of things that count as evidence that most people wouldn’t think of as evidence, but I’m really not sure where to draw the line.
It’s not so much from a “trusting someone” perspective but more from a philosophical perspective. Even with the current accepted trust levels on iNat there is plenty of room to not trust people (photos can be manipulated and dates/locations can be manually changed, so is someone saying that they saw a particular species but with no photo really that different?).
But anyway, I have a number of scenarios and in this case we will trust that they have been correctly recorded. I apologise in advance that there are so many of them, but I promise you that it’s interesting (well it is to me at least). The first ones may seem very obviously to be evidence, and the last couple may seem like obviously not evidence if you just think about them in isolation. But have a look at all the inbetweens. Where along the gradient do you stop?? I’m keen to hear what people think!


Is an organism evidence?

Is a photograph from a camera trap evidence? After all, nobody was there to see or hear the organism in person.

Is a museum specimen evidence? The location and date are certainly inaccurate if you just use when and where you took the photo of the specimen, but there’s still an organism there, and if you use the collection details then the location and date are correct.

Is a bone from a dead organism evidence? After all, it’s just part of the organism. What about a fossilised bone? It’s no longer organic. Sure, it’s not recent evidence, but is it evidence?

Is a dropped tail from a gecko evidence? The gecko is no longer here, but we know the tail must have been attached at some point.

Is a shed skin evidence? Again the organism is no longer here. What about a fallen leaf or branch?

Is a footprint or trackway evidence? After all, it’s not alive and never was. But what about a shell then? Shells aren’t alive, even though they are attached to the body. What about a fossilised footprint or impression?

Is a scat with remains in it evidence of both the animal that produced the scat, and the organism that it ate, the remains of which are visible in the scat?

What about a gall, or a feeding mark? We may not be able to identify exactly what species produced it, but is a swollen gall, chewn leaf, or scratched trunk evidence?

Is an illness or symptom evidence of the pathogen that caused it? Let’s just assume for a moment that we can accurately identify the pathogen based on a particular symptom. If I see an organism with that symptom, is that evidence of that pathogen?

Is a packet of chips (either type) evidence of a potato? It has been modified by humans, but there’s still a potato there. It has also been transported, possibly quite a long way, so you could say that the date and location are incorrect, but you cannot deny that there is a potato there.

Is a call from a bird or cricket evidence? No organism has been seen. If we could accurately record and depict a smell or taste on iNat, would that be evidence as well?

Is a photograph of a photograph evidence? The location and date may no longer be accurate, but surely it is still evidence that an organism was seen and photographed by someone. What if I have only a physical copy of a photograph that I want to upload to iNat, and to do so I take a photo of the photo and upload? What about a recording of a recording of a call?

Is a photograph of any sort evidence of humans? After all, photographs are a human invention. But ironically we may not necessarily see anything of humans or their artefacts within the photo. Is the very existence of a sighting on iNat evidence of humans who made the sightings and maintain the website?

I was privileged enough to watch a male lyrebird displaying recently, in which he imitated the calls of a number of different local native birds, including a kookaburra and a magpie. There were many of those birds in the area, but is his call by itself evidence of a kookaburra and a magpie?? After all, he cannot have imitated them if he did not hear them. He cannot make up a call like a human might, so we can trust that he did actually hear one, so how is it different from a human recording?

Vervet monkeys have specific alarm calls for specific predators. If a vervet monkey makes an alarm call that means “leopard”, is that evidence of a leopard? Yes, the monkeys could be wrong, but let’s assume for a moment that they are correct and did see a leopard. Would a recording of this alarm call be evidence of a leopard?

If a troop of chimps see a predator and start reacting to it - i.e. throwing things, trying to chase it away - but I don’t see the predator myself, is their behaviour evidence of the predator? I might not be able to identify it, but there is still an animal there that they are reacting to. What about a deer running away from a predator, but again I do not see the predator myself? Sure, maybe the deer was reacting to a noise it heard and there wasn’t actually a predator, but let’s pretend for the moment that the deer is correct and there is a predator. Does that count?

What about symbiotes? Is an observation of a healthy reef coral evidence of zooxanthellae? The coral would be bleached and discoloured without them, even if we cannot directly see individual zoothanthellae.

Land plants rely on mycorrhyzal fungi in their roots to help them extract water and nutrients from the soil. Without them, the plant would be dead. So is an observation of a strong, healthy plant evidence of mycorrhyzal fungi? We cannot directly see them, but without the plant would be dead if they weren’t there.

Is a fallen leaf or branch then evidence of mycorrhyzal fungi?? The plant could not have been strong enough to grow a leaf or branch without mycorrhyzal fungi, even if the plant may now be dead because the mycorrhyzal fungi have disappeared for some reason or other. After all, we say that a shed skin is evidence of an organism even if the organism itself is long gone and potentially dead.

Is then a large bleached or dead zooxanthellate coral evidence of zooxanthellae? They are no longer here, but for the coral to have grown this large it must have had zooxanthellae to help it.

Is a ripe fig evidence of fig wasps? The fig would be sterile and unripe without them, so to ripen it must have had wasps inside it. Equally, is an observation of the wasp then evidence of the fig? Many fig wasps are species-specific and can only emerge as an adult from a particular type of fig. If I see an adult wasp, it must have come out of a fig.

What about the yucca moth and yucca plant? They cannot live and reproduce without each other. The yucca plant is the sole food for the caterpillar of the yucca moth, so to see a large caterpillar (or adult moth for that matter), there must also be a yucca plant. And equally again, if I see a yucca plant is that evidence of a yucca moth? The pollination of the yucca plant requires a yucca moth, so does seeing a plant count as evidence for a moth? We know that a moth has been here, because otherwise there would be no plant.

The stick insect Parapodacanthus hasenpuschorum feeds exclusively on plants from the family Rutaeceae - is seeing an adult stick insect then evidence for a plant from the family Rutaeceae? Sure, the stick insect may have travelled a long way (they are very mobile and can fly quite well), but there still must have been a plant somewhere. You could argue that the location is not accurate, but when you record a sighting of something like a shed skin, you don’t say that the location is inaccurate because the animal is no longer there. If a grasshopper sheds its skin and then a huge storm comes in, and the grasshopper gets blown 5km inland and the shedding gets blown 5km out to sea where I find it on the waves from a boat, the grasshopper is 10km away at this point and was never even here in the first place, but you wouldn’t say that the location is inaccurate because it’s the location of the evidence, not the location of the organism. If a palm frond washes up on your local beach but there are no palm trees within 100km, you don’t say that the location is incorrect because the tree it came from is at hundreds of kilometres away.

So stemming from that, is an observation of any sort of herbivorous organism evidence of some sort of plant? If I see an elephant, is that evidence of the plants that it must have eaten to have survived and grown up until this point? What about carnivores? If I see a humpback whale, is that evidence of krill? If I see a lion, is that evidence of the other animals that it has eaten?

Is seeing Angraecum sesquipedale evidence of the moth Xanthopan morganii? The plant cannot be pollinated without the moth, so even if this particular plant had not been pollinated, its parents must have been.

What about further back in time? Assuming that the theory about avocado dispersal is correct, is an avocado evidence of the organisms that it evolved to be dispersed by? They may be extinct, so maybe it is not recent evidence, but the avocado could not have evolved without them existing at some point. What about other evolutionary questions? Is a butterfly with eyespots evidence of a visual predator? Is a wasp-mimicking beetle evidence of wasps, and is it evidence of the predators that it is trying to fool? These may not represent recent evidence and in some cases the other organisms may not even exist any more, but the simple fact remains that these features would not exist if those other organisms had not existed in the first place.

Is a photograph of an adult human face evidence of Demodex? Young babies do not have any Demodex, but it’s pretty inconceivable that an adult would not have them on their skin. So is a photo of a face evidence of Demodex? They are certainly there, but we cannot see them, and if they weren’t there the photo would not look any different.

Is then a photo of almost anything evidence of bacteria? It is almost impossible to not take a photo of something that contains bacteria unless you are taking a photograph of space, but even then the camera itself certainly contains some sort of bacteria. But if the picture didn’t contain any bacteria, it wouldn’t look any different.


Plenty to think about for sure. Apologies that it’s so long and there are so many examples, but I just keep thinking about all of these crazy things and coming up with more! What do you all think? Which of these counts as evidence and which doesn’t? I’m really interested to see where everyone draws the line, because the scenarios mostly change so gradually. So what do you think, and importantly why do you think that? Why are some things evidence and others are not??

4 Likes

Humans are not necessarily evidence of Demodex. Young babies don’t have them, and it’s possible for older people to get rid of them. We had to deal with that because one of the family developed allergies. Of course, I’m not sure how long they stay gone after being killed off.

1 Like