This step was performed on the blue book.
Then it was reversed.
The original intent of the experimental approach to bookbinding was to compare the flexibility of PVA and wheat paste on the spines of tape-bound books. Fellow binder Chris Busta-Peck suggested I try making the spine with wheat paste rather than PVA. The idea was that wheat paste and rice paper together would create a sufficiently flexible spine.
Wheat paste on its own would not be enough. It dries very hard, and a spine needs a flexible glue. (Most binders use either PVA or hide glue). Chris' suggestion was to use the wheat paste to glue rice paper to the backs of the signatures. The wheat paste would stick to the rice paper, and the rice paper would do the bending.
So I went through all the steps of rounding and backing the blue book with wheat paste in place of PVA. Before I created the hollow back, I pasted a strip of rice paper to the backs of the signatures and rubbed it down well with a bone folder. I then went through the entire spine construction process, again using wheat paste rather than PVA.
Because wheat paste takes longer to dry than PVA, I left the book for an extra day before even trying to open it. When I did, my worst fears were realised. The entire hollow back pulled away from the spine of the book.
So what went wrong? Why did wheat paste work for Chris but not for me? Well, he has been making a different type of spine than I have. Rather than rounding and backing his books, he's been squaring up the book blocks after sewing, then making the spines with flat backs. This means that his spines have less travel in them. When you open one of them, the spine goes from flat to concave. In a rounded and backed book, the spine starts out convex, goes flat, then goes concave when opened. This is simply more than wheat paste can do.
It was clear that there was no point continuing the bind with the spine in that condition. It practically fell off in my hands, leaving me with the naked book block. However, the failure of this part of the experiment was a classic "opportunity" rather than "problem". I had realised, partway through the experiment, that there was one more technique that I wanted to include: casing in as opposed to constructing the hollow back on the book.
So I re-backed the blue book with PVA, then cased it in.
Note, however, that wheat paste is still a vital part of bookbinding - I use it to glue up split boards, and it is the best adhesive I have found for leather.
2005: I have concluded that the debate between wheat paste users and PVA users is a transatlantic one. British binders use a lot of PVA, and there is a ready supply of reversible and pH-neutral stuff here. American and Canadian binders dislike it, and have learned how to make wheat paste serve their purposes. I don't know what the European binders do. Some people use a mix of the two - kind of a mid-Atlantic compromise. I am British in this regard, looking on PVA as the natural successor to hide glue, completely separate from wheat paste.