In college I thought I had a lot of books, and it's true that I had more books in my dorm room than most people: I'd brought most of my personal library with me when I moved across the country. But what was really remarkable about my library was how many reference books I owned, a collection that continued to grow for years until suddenly it didn't.
I started college with The Oxford Companion to English Literature, Benét's Reader's Encyclopedia, at least two dictionaries of quotations and three dictionaries of etymology. I had atlases and almanacs and style manuals, not to mention the NYPL Desk Reference, which I consulted so often that my roommate would cry, "here comes the nipple!" whenever I took it from the shelf. Over the next four years I added to this collection. I discovered The Princeton Encyclopedia of Poetry and Poetics at a used bookstore and figured I might need it. I bought Random House's Historical Dictionary of American Slang because I knew I'd need it.
At some point I became obsessed with the idea of getting the complete 20-volume Oxford English Dictionary--nevermind that I lived in studio apartments until I was past thirty. When I was working my law firm job I considered saving up $1,000 to buy it new. Once I started grad school I looked longingly at the used sets that would occasionally pop up in local bookstores. Eventually I compromised and bought a 20-year-old "compact" version--the kind with the magnifying glass--which I lugged home through the rain, more than a mile across town, so eager was I to have it in my possession.
And I used these books all the time. My first year of grad school one of my professors gave us a assignment that consisted of a long list of terms, titles, and names from the period, none of which we'd discussed, and set us loose to identify them and their relevance. This was in 1999. Google didn't exist, Wikipedia didn't exist, and I didn't have internet in my apartment anyway. I was able to sketch out at least preliminary identifications for some 70% of the entries from my reference books. (Most of the others I got from the library's Dictionary of National Biography, which I loved so much that I immediately wanted to buy my own set; alas, the cost was prohibitive.)
But at some point I stopped using many of these books. I had online access to the OED, which was faster and more current. I had Google and Wikipedia for when I wanted to know the population of St. Louis, or what year Charlemagne died, or why a quotation sounded so familiar. And I had real books--my growing scholarly library--for the more precise and detailed things I needed to know about the authors and texts in which I was slowly becoming an expert. During my past two moves I donated or discarded many of the reference books I once had, and the ones I've kept I don't use very often.
The exceptions are those books on subjects that my work touches on but that lie outside my field of immediate expertise. I no longer need The Oxford Companion to English Literature, but I sure do need their companions to the Bible and Classical literature. I need my encyclopedias of music and church history. Those are subjects in which I'm still a beginner (and often don't have better ideas about where to start when I need to brush up quick), but about which I need more than just fun factoids or whatever crap the internet might turn up.
But I guess it's not true that I have fewer reference books within my field than I used to; they're just different. Where once I prized encyclopedias and dictionaries and manuals, I now own concordances and variora, biographies and Complete Works. All the books in my campus office, all the books in my home office, indeed any book I wouldn't read at the beach or on the elliptical is, in some sense, a reference book.
But I still kinda want that 20-volume OED.