This study investigated the potential of students' written and oral questions both as an epistemic probe and heuristic for initiating collaborative argumentation in science. Four classes of students, aged 12-14 years from two countries, were asked to discuss which of two graphs best represented the change in temperature as ice was heated to steam. The discussion was initiated by asking questions about the phenomenon. Working in groups (with members who had differing viewpoints) and guided by a set of question prompts, an argument sheet, and an argument diagram, students discussed contrasting arguments. One group of students from each class was audiotaped. The number of questions written, the concepts addressed, and the quality of written arguments were then scored. A positive correlation between these factors was found. Discourse analysis showed that the initial focus on questions prompted students to articulate their puzzlement; make explicit their claims and (mis)conceptions; identify and relate relevant key concepts; construct explanations; and consider alternative propositions when their ideas were challenged. Productive argumentation was characterized by students' questions which focused on key ideas of inquiry, a variety of scientific concepts, and which made explicit reference to the structural components of an argument. These findings suggest that supporting students in productive discourse is aided by scaffolding student questioning, teaching the criteria for a good argument, and providing a structure that helps them to organize and verbalize their arguments. (Contains 5 tables and 2 figures.)