Closed access damages peer-review

I talked today with a scientist (R) whom I meet frequently and who works in a leading bioscientific research establishment (not a University, but with Nobel laureate and FRS on the staff). In addition to their day job, R acts as a peer-reviewer for a leader bioscience journal (J). R gets about 1 paper a week from J and has to return a review within 21 days. R does not get paid for their reviewing, which can certainly run into hours per journal. Like many other scientists R does it because it contributes to science. R and J take review seriously and J has given R training in how to review. The reviewing can be seen as a credit on R’s CV.
Some of you may not be familiar with how peer-review in scholarly journals works, so here’s a rough overview. An author  (A) sends a manuscript (M) to an editor (E) on J which decides on which reviewers (R, R1, R2) should review M. R, R1… are normally asked to decide on:

• whether M is in scope for J
• whether it reports a significant scientific advance (i.e. not repeating work already known)
• whether the science – as reported in M – is sound and potentially capable of being reproduced
• whether A shows that they are aware of published work that impinges on M
• and in many journals to give some idea of the “score” or “importance” of M. This could be “very important”, “minor advance, but useful”, etc.

R is normally dependent on:

• the material in M.
• the references (citations) in M to other work, C1, C2…
• R’s knowledge of the field from meetings, conversations, reading, etc.

R must keep M confidential but is normally allowed to consult close colleagues in confidence on small matters of fact.
R relies heavily on the material in the references (C1, C2…). These can contain background material, precise recipes, data, closely argued positions, etc. Without C1, C2 it is normally impossible to review the paper.
R tells me that on average there are at least 3 references (C1,C2,C3) per manuscript (M) which are closed and to which R’s institution does not have access. R cannot review the paper responsibly without C1, C2, C3. So what should R do?

• send M back and say they cannot review it
• pay the cost (3 * USD30 = 90 USD) for access to these references
• ask R’s institution to pay for access (why should they?)
• get an interlibrary loan (takes days and anyway costs money)
• ask a friend (me) for a copy of C1, C2, because Cambridge subscribes to these closed journals. I have to say no, because that would be a breach of copyright and whatever byzantine conditions the publishers have agreed with our library.
• do a bad review

THE LACK OF ACCESS TO PUBLICATIONS HARMS PEER-REVIEW
This is a sample of one, but by mathematical induction it applies to all peer-review. Therefore the hidden cost to science of closed access is enormous. Either the reviewers pay 90 USD per paper (which I strongly doubt), or they violate copyright (unthinkable), or they do bad reviews (certainly not) or …
So, closed access publishers, by preventing reviewers (and there are tens of thousands) reading your publications you are damaging peer-review. In a paper age this was accepted – you couldn’t read everything – or it took ages. But in the electronic age it isn’t necessary. We ought to be getting better quicker peer-review because of e-paper.
Are we? and do you care enough to do something about it? If not Open Access (the obvious answer), WHAT?