A good match, perhaps from the Zürich area...

[MarcoP]

I know I like to play Devil's Advocate, but I still think the a priori assumption that the marginalia consist of valid words or phrases has little ground. This does not mean we should not try to make sense of them, maybe they consist of exotic spelling and abbreviations, but we must not fall into the trap of the countless Voynich 'Theorists' who perform mental gymnastics to make 'sense' of a vord or line of VM text. The main issue with marginalia is the small sample size.

Though the marginalia are not entirely readable and do not make sense grammatically, we can say a posteriori that about 50% of the You are not allowed to view links. Register or Login to view. text is readable. Longish words like “poxleber”, “anchiton”, “dabas”, “multos”, “portas”, “maria” and (famously, see Panofsky) the fragment “so nim” are documented and perfectly legitimate (as well as the shorter “cere/tere” and “vix”). Of course, some readings are more ambiguous than others, but the quantity of readable text is significant. I am not aware of any “theorist” directly extracting so many legitimate words from four lines of text with a simple substitution (i.e. without resorting to heavy distortion and taking a lot of liberties).