Linda mentioned some "funky" results. These no doubt are due to the non-Latin1 Unicode punctuation characters in the citations. But a bit disappointing is that the # symbols seems to be stripped from the CSV file. For example, see Hurley-1605, which contains the citation:
''Nellis-Nelles Immigrants From the Palatinate 1710'', Vol. 2, published in 2004 by The Nellis and Nelles Family Associations and the Herkimer County Historical Society #2336b page 28
The # is missing from #2336b. This character would be helpful for extracting the ID (2336b) from the citation.