I’ve noticed this issue quite a bit on iNaturalist. Many observations in my country show exotic animals that have ended up in the wild somehow. Some of these animals are alive and well, and appear to be surviving just fine in the foreign ecosystem. Yet I see a number of users marking these animals as “captive,” even when they are not.
This practice harms the accuracy of records on introduced exotic species worldwide. Such sightings are scientifically valuable, especially if stable populations eventually establish because they provide both a location and an approximate date of when these species were first documented. They are effectively obstructing the pursuit of knowledge by dismissing these observations.
Personally, I also enjoy browsing unusual sightings, but people mislabeling them as captive makes this so much harder. Interesting records get buried among pet observations and other casual uploads, which then reduces their visibility and usefulness. It’s also quite frustrating to see a fantastic observation buried for seemingly no reason.
So, if you come across users doing this, please help by reminding them: if an animal is surviving in the wild without human care, it should not be marked as captive.
I’ve had similar issues with plants recently, when I have posted ‘garden varieties’ of plants that have self-set and are growing wild on roadside verges etc; people automatically mark them as cultivated. I broadly agree with your point, although there are endless grey areas in the wild vs captive debate.
Are you, as your wording suggests, arguing in favour of even individual escaped animals (or plants) that have no hope of breeding being considered as wild? Or do you subscribe to the notion that for an exotic species to be considered wild, it has to have established itself such that the species can sustain itself and produce new wild generations without human support?
A feral animal is one that has escaped from a domestic or captive status and is living more or less as a wild animal, or one that is descended from such animals.
By those definitions, an escaped animal can become feral and therefore by the iNat Help page it has become wild.
On the other hand, monkeys seem to escape from zoos and research facilities. While the owners or authorities are trying to recapture them, I would not consider them wild.
Some people mark such observations as not wild because they are concerned that escaped individuals that do not represent established (breeding) populations will mess up range maps and datasets.
While I do understand the reasoning of doing so, it is not consistent with iNat’s (admittedly rather idiosyncratic and non-intuitive) definition of wild vs. captive/cultivated.
There are a number of global and regional projects for escaped animals and escaped garden plants. Sometimes it can be helpful to add observations to such projects.
Sometimes the problem merely comes from the choice of words displayed to the users.
For example, when entering an observation on the Android app, the words: “Is it captive or cultivated" are used. But on the web, the words: “Organism is wild” are used. A reasonable person can answer “No” to both questions when referring to an escaped animal. An interesting experiment would be to change the words on the web to match the words from the Android app. (I don’t know what it says on the iPhone app.)
Yes, I know what the help pages say, but a small minority of users will look at those pages.
Of course, this is just a discussion about the words used in the English interfaces. I don’t know what confusion might be caused in other languages.
I suspect if we had 3 choices instead of 2 there would be less confusion: “wild”, “escaped”, “captive”. But such a change is not going to happen.
So for now, all we can do is what was said in the last paragraph of the initial post:
It’s a difficult subject with many different opinions. I personally think the current inat guidelines allow for distribution maps to not be accurate, especially when it comes to insects. Every now and then I see an observation of a butterfly or caterpillar in, let’s say Europe, even though that tropical species is native to South America or Southwest Asia. Many of these observations are close to butterfly houses or garden centres. I mark these ‘captive’ so they don’t mess up the distribution map, as these individuals clearly did not come all the way from the other side of the world on their own. Yet others will then mark them ‘wild’ because the insects “have escaped and thus aren’t in captivity” and “no one intended for the insect to be right there; it flew or crawled there on its own from the butterfly house down the street”.
This way of thinking truly blows my mind, but in the end it’s correct according to the guidelines.
Those observations can be really important for documenting the arrival and spread of non-native and potentially invasive species, and it’s strongly recommended that they be listed as ‘wild’. Especially with things with short generation times like insects as that individual may not have been one that escaped, but the wild born offspring of ones that did a few generations back.
The distribution maps are only very rough guidelines and many are very inaccurate even for well documented native species in their own home ranges.
Distribution maps need to follow the observations (assuming the ID is accurate and correct), not vice-versa.
Yes, that all makes sense and I understand that inat observations can be important for documenting the arrival and spread of non-native species. But I just can’t see how a tropical butterfly flying just outside of the butterfly house in Europe that it clearly just escaped from, is ‘wild’. Especially if that butterfly couldn’t possibly survive the cold winter and/or because the host plant is not even around.
That data is then also exported to a website like GBIF, and along the way we get information like “species X occurs in the tropical rain forests of Colombia and Ecuador, and in Germany”. But that’s the problem: the species doesn’t occur there and never reached that place on its own; only one individual was seen there after it had escaped from a butterfly farm that it was brought to by humans.
I once tried arguing this POV with the example of a scale insect on an orange fruit I got from the grocery store in a snowy wintery climate where there are zero orange trees. But the groupthink on iNat was convinced this must be a “wild” observation, because humans didn’t knowingly transport the insect, and because “this is how we can document invasive species”. Never mind that this is literally impossible for the scale insect in question. Yes, that documentation is sometimes useful, but people can be way too literal with these guidelines sometimes.
I would have no problem with such observations not becoming RG and not being shared with GBIF.
But as an observer it is fairly annoying when one has observed an escaped organism in good faith – i.e., when one has followed iNat’s guidelines about what qualifies as wild and non-wild and has chosen to document that particular organism precisely because it is an unusual data point (an organism out of place) – to find that one’s observation has been marked as “not wild” and thus heaped together with pampered pets and zoo animals and other observations which are of little interest for iNat and which one has chosen not to observe. An escaped organism is different in important ways from one that is under the care of humans. It is demonstrating agency and the limitations of human control – regardless of whether or not it will survive to reproduce.
For all purposes I think it’s massively better to have one consistently applied definition than dozens of competing ones. If people are worried about data sanitation a much better use of their time would be to make sure things like the “escapee/non-established” observation field and establishment annotation (on applicapable taxa) are being applied correctly than getting into endless games of tug of war in the DQA or rehashing the same argument on the forum for the 30th time when it’s clear where the website policy stands.
I would consider your scale insect a waif with low probability of establishment. There’s no good category on iNat for assigning such organisms. If it were my record I’d probably call it captive for lack of a better category, assuming I thought it worth submitting. I think many records of waifs are worth reporting because some just might establish. But they’re not really wild by iNat standards.
Still, a non-enforceable policy is just as effective as no policy at all. On the other hand, a policy that is strictly adhered to by The Powers That Be is… “let the tug-of-war in the DQA decide!”
Forgive me for putting it so bluntly, but I feel that if I have to come down on the side of giving more latitude to those who find it a wee bit annoying that their observations get marked incorrectly sometimes, or those whose work depends on accurate data without spurious outliers in order to draw distribution maps and study wild populations, I think an abundantly cautious approach should favour the latter over the former. I say this as both an observer who has experienced the slight irritation of having my records of escaped plants wrongly marked as cultivated, and a researcher who has experienced the total absurdity of having range maps and geographic models skewed by cultivated plants that actually should have been marked as wild.
Which is why I suggested that ideally such observations would not become RG/not be shared with GBIF. I am not unsympathetic to the concerns of researchers.
De facto it is a moot point, because the number of users who are determined that such observations should not count as wild regardless of iNat’s guidelines means that they invariably become casual eventually (generally sooner rather than later).
And yet it is nonetheless also true that lumping escaped individuals together with captive ones erases important and interesting distinctions. (Yes, there are projects and observation fields that can be and are used to record this, but they are less intimately integrated into iNat’s interface.)
I do not see how an escaped tropical butterfly is really all that different from a feral pansy growing in a sidewalk crack (which many people would be willing to accept as wild). Neither is likely to survive for more than a few weeks. While the pansy may be capable of reproducing, I have never seen any that result in enduring populations in the conditions where they sprout.
I would also question whether such observations are as much of a problem for data integrity as they are made out to be. If someone concludes based on a handful of records of escaped individuals that a tropical butterfly is found in Germany, they are not doing their due diligence and checking what those records represent. Outliers in particular should always be examined before accepting them, and if the circumstances of the find are transparently documented by the observer, it is straightforward enough to omit it. I realize that of course this is additional work for the researcher, but they ought be vetting their data set anyway and obvious, clearly documented cases of escaped individuals strike me as fairly unproblematic compared to many other sorts of unwanted data points that one might end up with (garden plants that have not been marked as wild being a huge one).
As a researcher who has used geographic data from both iNat and non-iNat data (eg, VertNet, GBIF), I would vastly prefer organisms to be marked as “wild” for two reasons:
A consistent dataset is much easier to work with and can be used to produce accurate results. A dataset that is inconsistent in terms of how data is entered/defined is much harder to work with to produce accurate results. If I know what “wild” means in terms of iNat observations, I can make a clear decision on how I want to use that data. If wild/captive is applied idiosyncratically, there may be no way to use the data for some purposes other than checking every single observation (which is a lot of time/effort).
Every researcher worth their salt who is using iNat data should be doing some basic datacleaning/sanity checks. There are extralimital outliers in GBIF from museum data already (incorrectly recorded locations [I remember a subtropical frog from Norway in a historical specimen] and “waifs” [I remember a tropical frog from a Colorado supermarket]). These need to be addressed by the same processes by which one would check any iNat data as well. Having range outliers from iNat data may add more points to address, but doesn’t change the fundamental processes a researcher should use. If a researcher doesn’t due their due diligence for their dataset, that’s on them. Additionally, those extralimital datapoints often have value for some research applications. Digging them out of the sea of casual observations (when incorrectly marked) is much more work than screening them out of RG observations because they are obviously out of a species standard range (pretty straightforward and quick).