If people aren't checking these sources, or alternate sources, at all that means that I could invent a totally fictitios reference and have it accepted. That's scary.
Wikipedia is based heavily on trust and assuming good faith. I trust other editors not to invent references and I don't do it myself. But if a sufficiently sophisticated troll were to try it I imagine it could go unnoticed for a very long time indeed. Eventually an expert on the topic would stumble on the nonexisting reference, be puzzled and try without success to look it up.
And - to any aspiring trolls out there - *please don't try this*. Let's just assume you could, okay? I mean, even articles on fictitious topics can stand unnoticed on Wikipedia for years. I've found several examples in Norse mythology articles - articles on gods apparently invented out of whole cloth. I don't think that is a troll-driven phenomenon, presumably most of these stem from fantasy literature which someone has in good faith mistaken for authentic mythology.
Who would ever look in the edit history for a reference? I think it would have been an excellent reference to put at the bottom of the article. I have too many other things to occupy my time here without getting distracted by birds, but my inclination for fact-checking was to reach for Godfrey's "Birds of Canada". That would give a different perspective, and I could add that as a further reference. An American could give a view about the bird in Alaska; a Russian or Norwegian could also provide sources in those languages, and we would still have room for references from the Southern Hemisphere.
An important difference is that your bird book is in English, like the encyclopedia we're writing. In any case, if I thought that the observation on the arctic tern which I added from the book was somehow unique to Iceland I would have added the book as a reference. Since I assume it isn't I didn't.
If I had created a reference section and put my book there, alone, it would have been misleading since it would imply that the article was written using this book as a source - not so, I only added one sentence and didn't really fact-check the rest. It would also imply that it is for some reason appropriate to cite an Icelandic source here. That is not so, in my opinion, since good English sources are available.
Making something checkable, even through a rare reference, is the responsibility of the contributor. Actually checking it is the responsibility of the reader. If the reader does not fulfill his responsibility it's not your fault.
Well, I wouldn't exactly be very helpful by supplying a source in a language which the typical reader of the article doesn't understand.
And I disagree with the model of references as something intended to protect the writer from accusations of making things up. I see them primarily as an aid to the reader in finding more information.
And again I'd like to emphasize that we probably agree more than we disagree on this. In most reference related situations we'd probably both do the same thing. I'm certainly never happy with an article until it has good references, even if it is short. Look at [[Lofn]] for an example.
Regards, Haukur