Ups, seems like the Google Books algorithm is not completely free from mistakes after all. While searching for occurrences of the word Jass in the 19th century (with the underlying intention, to investigate if this word could in any way be related to the word Jazz, which appeared at the beginning of the 20th century) I unexpectedly encountered an digitization error in the Google Books results!
One of the books listed in the search results was the College Journal of Medical Science.
But following the link, I discovered, that the word in the mentioned paragraph, was not actually jass but pass!
I have to admit, this is most likely due to the fact, that the word is actually printed very unclear. However, it does raise the question, of how much to trust the Google Ngram Viewer (which is based upon Google Books data), especially with queries involving older titles. I assume, that most of the time, these small errors can be flattened out, if there are enough other, correct, occurrences.
Nevertheless, I found the case interesting enough to mention here :).