>> "And how come the distribution of actually pronounced <-ur> in
>> Icelandic is almost identical to where <-r> in On is recorded?"
>
> One reason I posted my post was that I pointed out that there are
> many modern Scandinavian dialects, not only the Icelandic ones, that
> have preserved the 'r' ending in perfect shape (at least for the
> adjectives and to some extent strong masculine names such as:

When I wrote the above I was not aware of the various wacky Swedish
dialects out there. Thanks to the efforts of a certain Swede this is
no longer so :)

> then it could have been just an odd, local mutation. But even today
> in modern Swedish we may very well say 'en långer dag' (ON langr
> dagr), 'en blåer båt' (blár bátr), even though it is a bit dialectal
> and illegal in formal written Swedish.

I wonder if this is in some part due to German influence.

> Thus, the silent r theory must be bogus, of course.

Certainly. I suspect it has come about from people seeing different
versions of the same name with and without the 'r'-ending and assuming
that it doesn't matter. Maybe you see 'Haraldr' half the time and 'Harald'
half the time and you only ever hear it pronounced without the 'r'.
Assuming that the 'r' is "silent" may have seemed like a good guess.

Kveðja,
Haukur