Jeff, could you tell us why they are not consistent and what sources to look for more accurate notation?

I wonder, once again, how much of this is due to tradition being in place, but places where they were not normally observed as they should have been in tradition they were marked?

Like opera.

It is my observation that in the sources not all places where there is a conjunction of two consonants actually receive liquescents. The examples you give for the First Sunday in Advent do not have liquescent neumes in the St. Gall notation in the Graduale Triplex. I have heard some advocate that when there is the conjunction of two consonants or vowels, you should always make liquescence, whether it is in the notes or not. But then, why are so many such places not provided with liquescent neumes in the staffless notation? I think, rather, that it is a stylistic matter, and that just a smattering of liquescents is attractive, though a systematic use of them would be overdone.

Seriously, the manuscripts were not made to be published and used worldwide. Any choral conductor knows that certain rooms are going to require more time on certain consonants than in another acoustic. To my ear, there's nothing more unmusical than a mechanical rendering of liquescents as if the notation is the music. If you sing well, and pronounce well, the liquescent phenomenon will take care of itself. Thanks to Jeff O. for pointing out that the notation does not convey every detail of performance practice, or in a consistent fashion.

Also interesting to note that in many cases double vowels such as in "tuum" were likely sung as a single vowel, since they are notated in the manuscripts as one neume. We so often hear instead "tu-um" (or worse, "tu-wum"). If we can't go from one vowel to the same vowel with a legato technique, how can we possibly sing "mea" or "tuo" which contain two different vowels?