With
leið,
laug and
kló I can assume you're referring to nouns of the feminine gender, right? This is quite tricky because as far as I am aware this started to drop away even when the earliest records began. It's definitely an archaic factor in Modern Icelandic now. You can see it in other feminine nouns of the same class (
stund - 'time/moment') where in a normal usage, you'd just have dative
stund but if you wanted to use an expression such as '
to leave something until the last moment' then it'd be
að draga eitthvað fram á síðustu stundu and lo-and-behold we get the old dative-u ending. It's a really unclear system and one I don't imagine speakers really have(had) a good hold on and it would therefore have been prime to just drop away.
Now, that's how the situation is today but it could very well have been
more regular in the past. A giveaway is the rounded vowels that exist in this class, which rounded the stem vowels and umlauted them and then dropped away. That's a good signal for a
u-ending but even in Norse for a lot of words that had dropped away. I think for a lot of these words you're looking more at Proto-Norse when the rounding took place because of the dative
u-ending. Tricky issue is, these nouns also had a nominative
u-ending which also probably clouded the issue of rounding the vowels.
The ing/ung endings are pretty resilient actually. Now in Modern Icelandic, the dative -u remained and actually spread over to the accusative where it wasn't historically there. In Norse, if you have a disyllabic stem ending in ing/ung then my bet is you will find a
u-ending in the dative. Monosyllables are more difficult, especially if they don't contain vowels that would have been rounded. I believe these often showed mixed paradigms, occasionally with the ending and occasionally without, as it was a pretty muddy system to apply any sort of analytic technique to see when it would need to be there or not.
There was a similar thing that happened to the -i ending in masculine singular dative. The system wasn't clear and then a pattern emerged where long vowels and singular consonants shifted to an i-less paradigm while complex clusters retained or added them. There was just an inkling in some cases of what the rule seemed to be, and then that in turn completely shook up the nouns and effectively made it a rule (absolutely full to the brim with exceptions though
).
Historically, I think it was there for all the nouns in the class but then started to drop away (i.e. when not with disyllabic nouns in that class - or ones that didn't end in -ing/-ung).
