I would agree that with your postulate that if we agree there are definitely adaptive changes within a species that over extremely long periods of time they might add up significantly. I'm not sure that the vastness of the fossil record alone clearly shows that the transitional forms offers us a slam dunk that proves those adaptive in species changes explains the plethora of species we have on hand...
It's worth remembering that
all creatures are transitional, in as much as evolution is a continuous process.
In investigating transitional species, one selects representatives of the distinguishable start and end points one is interested in (e.g. non-flying to/from flying, land-based to/from water-based, brachiating to/from bipedal, etc.), and then looks for creatures that have phenotypes showing intermediate or transitional features. So they're transitional with respect to the particular features under consideration.
... I don't think evolutionary science is anywhere near the point we have progressed with physics, chemistry or astronomy. Perhaps the brightest tool for panning this out will be the DNA research but it may turn out to be a double edged sword. Empirical evidence I believe is lacking.
Molecular genetics (DNA research) has already given evolutionary science a statistical rigour approaching that of the other physical sciences.
Take the finches that I mentioned with respect to the series of changes you potulate would happen over time. During the drought season it is true they adapted with beak size changes to their environment but as soon as el-nono brought back a series of rainy seasons their beak sizes returned to normal.
This is an effect known as 'Sisyphean evolution' which probably makes the finches a less than ideal example of radiative evolution; however, there are hundreds, if not thousands, of classic examples of radiative evolution in the Galapagos, that don't have this constraint; in birds, mammals, insects, reptiles, plants, etc. (I just watched a couple of documentaries on the Galapagos by David Attenborough).
So the question is whether there always exists a upward trend of "positive" changes that make up a series that would lead to transitional forms.
It's not clear what you mean by an 'upward' trend of 'positive' changes, but I hope it was simply expressing a clear trend towards features typical of the chosen transition end-point exemplar, rather than a value judgement.