Yeah, it definitely seems to be a recording artifact, easily fixed if anyone wants to. Also, what King mentions is a good idea - crossfading which could be done on a track basis as well as within the instrument.
I generally think that there is something about the sound of both cellos and violas in their forte dynamic that is extremely hard to capture with a completely smooth, natural sound (especially with infinitesimally short legato samples) because of the complex harmonics involved. Cecil Forsyth pointed out in his great text on orchestration that the viola is too small for its range - which is exactly why it is such an interesting contrast to the cello, which is perfectly sized for its range, and which also accounts for the beautiful timber it has. With those complexities of harmonic content, one is apt to encounter immense diificulties in recording every possible variation of sound. And that is what the VSL has ambitiously attempted.
On the other hand the higher and lower legatos of both violas and cellos are fantastic, not to mention the legato violins which in all dynamics are probably as close to perfect as can be done.
(The only reason I don't mention other instruments is because I haven't gotten to them yet - I am studying these fantastic sounds IN DEPTH.)
By the way - and I know this is a ridiculously demanding and pedantic criticism - I have noticed also that there are VERY SLIGHT problems with the velocity on various string instruments' legato transitions. You have to do some very careful adjustments of the initial note velocity vs. the legato-connected note's velocity to get a natural effect in some cases, because otherwise the transitional legato sample starts to sound almost like a new attack. Of course, that is no problem to do in your sequencer, and to ask the VSL to correct for this is like asking for the moon with they already have given you the stars. (To quote a famous old Hollywood movie whose title I won't mention.)