Revisiting the Shift from the Roll to the Codex Book Form

Recently, Brent Nongbri published an article in JSNT raising valid concerns with regard to using objectively dated rolls to assign dates to codices based on palaeography (see previous post). This JSNT article, and the subsequent interchange with Nongbri on the blog, got me thinking about the fantastic new resource put together by Grant Edwards.

"The Collaborative Database of Dateable Greek Bookhands"

The online database is very user friendly and it's sorting and filtering tools are quite intuitive. One drawback about using the data found in LDAB (which is a fantastic resource) is that many of the dates for the material referenced are not derived objectively.  Thus, any conclusions relying on the dates of the material obtained from LDAB (for example determining when the codex supplanted the roll) may not be as secure.

I thought that it might prove to be an interesting exercise to query the new "Collaborative Database of Dateable Greek Bookhands" for this information. Of course, objectively dated manuscripts are extremely rare and one should be careful about drawing sweeping conclusions from the information. 

All the same, using the "Filter" option to show only the format of "codex" gave only 32 results. The earliest codex being P.Bodmer 20 of the 4th century with some examples extending out to the 9th century (though with only one other 4th century example).

Using the "Filter" option again to show only the format of "roll" yielded 68 hits. Of  these, only 8 rolls were dated from the 3/4th century out to the 8th, with the bulk falling somwhere in the 1st-3rd century mark. 
I thought it intriguing that these more objectively dated results seem to support the well noted trend, that is, that the 4th century saw a shift in book formatting from the roll to the codex.