The "them" in the sentence you quoted refers to "the original public records".
The locks you refer to were around the new, value-added work: the digitized, processed, cataloged version of the papers they had made by adding value to the original public-domain documents. The locks were not applied to the original documents.
It's like the difference between storing a touched-up version you made of an Apollo 11 photo in a vault, and stealing the original photo negative from NASA and putting that in a vault.
JSTOR doesn't claim that its documents are public-domain:
Our licenses from publishers are non-exclusive, meaning that the publishers are free to license their content to others to digitize or make it available in any way they might wish. [0]
Not only were many of the papers in JSTOR funded by the public, but JSTOR itself claims to be funded through by the public through universities:
We do this with funds provided by thousands of libraries and institutions, all of whom are our partners in disseminating access around the globe. [0]
JSTOR cannot magically put documents in the public domain. It needs the permission of the original publishers. JSTOR is responsible for scanning a whole ton of old journals which would otherwise only be available as physical copies in libraries. Nothing JSTOR has ever done has ever made anything less accessible than it was before.
How is creating network and identity based locks around the papers not placing new locks?