This study examines first graders' epistemic reasoning, in tacit "practical epistemologies" reflected in thinking about an investigation of their own design. I analyzed children's epistemic reasoning, following a design experiment scaffolding increasing regulation of scientific inquiry in a domain they studied in depth. Participants consisted of children in two first grade classes (n = 35). The study capitalizes on the final phase of the design experiment, wherein pairs assumed substantial responsibility for design and implementation of an investigation pursuing their own question. I conducted structured interviews with each pair, eliciting their thinking about their investigation. Forty-three percent of the children attributed uncertainty to their findings. Forty-nine percent transcended phenomenon-based reasoning in thinking about how to improve their study, reasoning instead in terms of relation between variables. Forty percent posited additional potentially influential variables, reflecting a tacit understanding of the lack of any straightforward one-to-one correspondence between cause and effect. These findings suggest that epistemic reasoning is plastic, sensitive to instructional opportunities. I argue that the children's relatively deep familiarity with the domain, experience in assuming increasing responsibility for structuring scientific inquiry, and investment they had in their own investigations contributed to the surprisingly robust epistemic reasoning they revealed in this context. (Contains 3 tables, 11 figures, and 10 footnotes.)