Photographs have always played a major role in geographical studies. Ever since the invention of photography in the 1830s, it has been assumed that photographs are perfectly suited to help answer that eminently geographical question, "what is this place like?". Photographs can convey a great deal of information about the appearance of a place far more succinctly than words. However, more recently, some geographers have begun to treat photographs rather differently. Instead of using them as descriptive illustrations that simply show what a location looked like when the shutter snapped, increasing numbers of geographers are thinking about the ways that photographs can be active players in the construction of a range of different kinds of geographical knowledge. That is to say, instead of thinking of photos as transparent windows that allow one to peer into places he would never otherwise see, some geographers are starting to think of photos more as prisms that refract what can be seen in quite particular ways. And still others are emphasizing the importance of how photographs are seen and used by particular viewers in particular contexts. This essay focuses on some of the implications of this new work for using photographs as illustrations in human geography. This interest in photographs by a range of different geographers raises some pressing questions in relation to the use of photos as illustrations. The author explores what some of these questions are and discusses five rather different ways of approaching photographs. It is less a "how-to-do-it" lesson in taking good photographs, and more a "why-do-it" and a "with-what-consequences" discussion of using photos in one's work.