Howard,
Couldn't agree more with your very fair post of 8.41 (before the drifts started again). Steve's number crunching is mine and the 90% is mine. It applies to that material, as stated several times above, which was collected and published in the period c1890-WWII, largely in southern England, so a corpus of about a couple of thousand songs, not including one-offs which may or may not have been folksongs. Actually about 10 years ago when I did the survey the more precise percentage was 89%, but we have discovered a few more since then. If you want the full description of the material studied it is in the intro to the new edition of The Wanton Seed.