The rolling window analysis of thorn and eth shows that lines 1924-2138 of the Anglo-Saxon poem Genesis A had a written, Old English source.
It has long been known that at least lines 1982-2005 and perhaps 2039-2095 are not drawn from the Latin Bible like the rest of Genesis A. But I hadn't realized the implications of my own data: that only a written source in Old English could account for the anomaly in the thorn/eth ratio at that part of the poem.