I can't quite imagine what it would have been like to conduct a literature search before the age of the Internet. It must have been a combination of browsing through hundreds of Tables of Contents and "treeing" through the references (they cite this paper, so I should look that one up too). Many hours would have been spent in the library stacks.
If I need to conduct a literature search, I first decide on which of many databases I will use. Google has its own Scholar function for searching through academic resources, which will pull up long lists of references on relatively obscure topics, complete with "Find It At *U" links for the vast majority of results (when I'm on campus, of course) so I can download the .pdf file.
Only once has the Internet failed me - that I'm aware of. It was a matter of more old-fashioned searching. A paper I found online cited a very relevant study, which I wanted to read for myself. It was from 1988, in a relatively obscure journal. This isn't necessarily a problem; I have a .pdf of an aritcle published in 1985, and 1989. But this article does not appear to exist on the Internet - the abstract is all over the place, of course, but the article itself has eluded me.
So it's off to the library for me. For just one of the 93 articles I have listed in my comps references, the article has been protected and *U has a print-only subscription. So I have to find the Science Stacks, and use the call number to find a specific shelf and a specific volume. I'll actually have to photocopy the article (or, heck, scan it and create my own .pdf). And while I'm there, I might spend a moment's silent contemplation in awe of how hard finding articles once was.