We examined the strengths and weaknesses of three diverse scroll control modalities for photo browsing on personal digital assistants (PDAs). This exploration covered nine alternatives in a design space that consisted of three visual interfaces and three control modalities. The three interfaces were a traditional thumbnail layout, a layout that placed a single picture on the screen at a time, and a hybrid that placed one large photo in the center of the display, while also displaying a row of neighboring thumbnails at the top and bottom of the screen. In a user experiment we paired each of these interfaces with each of the following three scroll control modalities: a jog dial, a squeeze sensor, and an on-screen control that was activated by tapping with a stylus. We offer a simple model that classifies our experiment’s interfaces by how much they provide visual context within the photo collection. The model also classifies the scroll modalities by how tightly they correlate scroll input actions to effects on the screen. Performance and attitudinal results from the user experiment are presented and discussed.