This is not particularly surprising. When the transaction cost to read an article is lower, there's less "need" to use the work you've uncovered. By analogy, Walter Murch says in his book, "In the Blink of an Eye" that film editors need to avoid "seeing around the edge of the frame" - when the editor knows the hard work that went into getting a particular shot to work, the 'sweat investment', it's harder to discard that shot because it feels like a waste -- even if that's not the best shot to use. Similarly, my investment into getting an article means I'm more likely to use it the harder it was to obtain. By decreasing that cost, writers can be more 'objective' about what's actually useful.
Second, because I can easily get a different article, if the one I have doesn't say quite what I need, I can find something else.
Leading to more herding and informational cascades (http://www.info-cascades.info/) wherein individuals cease to follow their own thoughts and instead rush to verify what they have heard through (in this case) the academic grapevine.
Not really. They complement each other. The one you showed claims (rightly so) that openness is a fundamental requisite for the progress of science. This one, shows how that openness results in more hits on the best papers on a given subject.
Think of it this way... even if you have easy access to 100x more papers than before, the number of citations that you put in your paper is still the same. This means that the you will try to pick the best references from the set you have available, which, in turn, will result in some particularly good papers being cited more often than they would otherwise.
Second, because I can easily get a different article, if the one I have doesn't say quite what I need, I can find something else.