Historical frequency of expression "and/or": Corpus search

What is the historical frequency of the expression “and/or”?

I have a feeling that I almost never see it in older texts, but that it is has become exponentially common in the past five or ten years.

However, Google’s Ngram Viewer doesn’t seem to work because of the slash.

Any suggestions? Or does anyone already have data on the historical frequency of “and/or”?

Using the Corpus of Historical American English (http://corpus.byu.edu/coha), it appears your assumptions are correct. That corpus tool doesn't allow for the embedding of the charts it generates, but here are the numbers:

1860s - .06 times per million
1870s - .11 times per million
1880s  through 1910s - not at all
1920s - .31 times per million
1930s - .61 times per million
1940s - 1.15 times per million
1950s - 3.46 times per million
1960s - 7.51 times per million
1970s - 6.42 times per million
1980s - 10.43 times per million
1990s - 10.59 times per million
2000s - 13.53 times per million

You should be able to visit the corpus yourself to do a keywords in context search (if you're interested in that). The search term there would be "and?or" (where the ? represents a single wildcard character ... this gets around the slash problem and relies on the fact that there is no English word or phrase with any character there except and/or).