In Rep. 2.16, I don't think Scotus is drawing a distinction (rational or formal) between bases and dispositions. Instead, I think he's drawing a distinction between different bases. He's saying something like this: "intellectual and voluntary activity have different bases. The basis for Intellectual activity is a part of the soul that we call 'intellect', and the basis for voluntary activity is a part of the soul that we call 'will'. But these are just two parts of the soul --- two things, so to speak, that are stuck together to make up the soul." It's almost as if we're talking about the right and left halves of the soul, except there are no right and left sides to the soul, since it's immaterial. But we are talking about absolute things/parts/constituents of the soul.
Of course, Scotus goes on to insist that these two parts are formally distinct from each other, seeing as they are "unitatively contained in the soul." But by that, I think Scotus just means that those two parts are joined so tightly together that the soul wouldn't survive if they were pulled apart.
So all of that distinction talk in the 2.15 passage is, as I read it, just talk about the distinction between the categorical bases for different activities. Nothing about a distinction between dispositions. It's a little tricky because Scotus _talks_ about powers rather than bases, but as he likes to point out, when he talks about powers in this context, he's really talking about absolute bases (or "immediate bases" I think is how he puts it in Rep. 2.16).
As for dispositions, I think Scotus rules them out as things or entities of any kind. On his view, there's just no such thing or entity or being as a "power" or "disposition." Of course, we can _ascribe_ powers and dispositions to things. We can say "I have the power to think, raise my arm, and whatnot," and all of that is true. But that's on the level of linguistic/propositional ascription. If we lay ascriptions aside and turn our attention to the things themselves, to the ontology, then we see a different picture.
On my reading, Scotus would say there's just the categorical basis, i.e., the absolute parts or constituents in the thing we are ascribing the power to. When we say "X has the power P to do Y," that's true if X has the right kind of absolute parts/constituents to do Y. But there's nothing that P would refer to over and above that. So an ascription like "I have the power to raise my arm" is true not because there's me, the right kinds of bones and muscles, and _some other thing/mode_ called a power. The ascription is true simply in virtue of me having the right kind of bones and muscles.
I think that's the point of Scotus criticizing first Giles/Aquinas and then Henry in the 2.16 passage. He first rules out the view of Giles/Aquinas. Their view is that powers/dispositions are things (qualities) that are really distinct from their absolute bases (soul). As Scotus sees it, powers/dispositions just can't be distinct things over and above their bases.
"Okay," one might think, "that's fine. So maybe powers/dispositions are somehow _the same_ as their bases. Rather than being distinct entities in their own right, maybe they are the same as their bases, but a different mode of them or something." That's the view of Henry: a power/disposition is really the same as its basis (call the basis X), just in a relative mode. So taken absolutely, X is just the absolute thing (a bit of heat, a substantial form, a lump of matter, or whatever). Taken relatively (with respect to the relevant activity), it is a power. Henry's view is similar to one that's somewhat popular today: all properties are _both_ categorical and dispositional in nature. Henry would probably fall into that camp: parts and constituents are both categorical (when considered in themselves) and dispositional (when considered with respect to the activities that are based on them).
But Scotus rules out that too. He says: no, a power/disposition can't even be a different _mode_ of the absolute part/constituent. On the contrary, it has to be just the absolute part/constituent. Nothing more. And that, I take it, essentially rules out the reality of dispositions altogether. Scotus is what we might today call a categorical realist: he thinks there really are powers (at the level of ascription), but on the ontological level, there's really nothing more than the categorical bases. To use a modern example: what is the fragility of a wine glass? A categorical realist would say it's really nothing more than the molecular structure: the molecules are made up and joined together in a certain way, and that's all it is to be fragile.
