Science writing at a university has to be one of the world's great jobs. If the institution is serious about its research, you're a kid in a candy store. In my case, at Ohio State University, with more than 3,500 faculty, the question is what to write about first—not where to look for stories. Big universities are that way, but the same rules apply for smaller places that are intent on doing great research. Let's begin with the basics. While public information officers at universities face a buffet of varying tasks—from covering boards of trustees' meetings to athletic scandals to student riots—the role of the science PIO is more focused: Concentrate on university research; explain what is new and why it is important to the public. Stated that way, the job seems simple, but science writers at a university may have to jump from astronomy to immunology to psychology to anthropology all in the same week. That represents a lot of intellectual gear shifting; but remember, the rules about reporting on research generally stay the same from field to field. What is the news? Why is it important? What is the context for the research? That is, what are the questions that drive it? Why should the readers care? And last, do the findings point us somewhere new? The only things that change from story to story are the researchers' language and the culture specific to their fields. Nearly every time I give a talk on university science writing—and there have been dozens—someone asks the classic question: How do you find your stories? The glib answer is “Everywhere;” but in truth, that's pretty accurate. Some people envision situations where top researchers have a “eureka” moment and then immediately get on the phone to the campus science writer to get the word out. Or perhaps the researcher's department chair or dean, ever attuned to their colleagues' work, is the one to pass along such news. I wish that were so; but sadly, it's more likely that researcher, department chair, or dean will never think about calling a writer until long after everything else is done.