Brian - those are for the selection she made for the 2 vol published collection. In the introduction she says:
...he noted nearly 3,300 tunes [footnote: In addition he noted nearly 1700 tunes in the Southern Appalachian Mountains...making a total of nearly 5,000 tunes]. Of these about 480 are dance tunes, 180 singing-games, and 170 chanteys, leaving about 2,470 songs and ballad-tunes.
It is from this last that the present selection is taken. It consists of 1165 tunes (413 separate titles) and though it numbers somewhat less than half the tunes that were noted it may nevertheless be said to represent the corpus of the collection
She goes on to categorize the omitted as: 1) Versions with only slight differences from included, 2) tunes only partly remembered, 3) songs of doubtful traditional provenance 4) a few noted from Irish immigrants in Marylebone W/H considered alien 5) a number of songs with commonplace tunes accompanied by texts which are corrupt, fragmentary, or of little interest (..mostly of the Come all ye type).
The figures from Conclusions seem to be Sharp's estimates from his whole collection. But since Conclusions was published in 1907 I don't know what proportion of his songs had been collected by then. It's possible the distribution he gave there might be altered in the full set of songs he collected (Karpeles gives 1903-1923, the year he died, as the dates he collected).