Does reading a mathematical proof for validation engender different behaviors from reading it for comprehension? Experts and novices each read two mathematical proofs under different sets of instructions: they were asked to understand one proof, and to assess the validity of the other. Their eye movements were recorded while they read and were analyzed to investigate possible differences in attention allocation, in cognitive demand and in the mathematical reading process. We found negligible differences in reading behaviors under the two sets of instructions, and we discuss the implications of this for theoretical development, research methodology and pedagogical practice.