The Virtual Ecological Research Assistant (VERA) is an interactive environment for supporting learning about the domain of ecology as well as the process of scientific modeling. The cognitive theory of learning in the VERA experiment is based on two main constructs: science makes progress through model construction, evaluation and revision, and inquiry-driven modeling enhances learning about science. The design of the VERA technology, too, is based on two main constructs: a virtual laboratory that grounds construction of conceptual models in experimentation, and an artificial intelligence compiler that automatically generates simulations for evaluating the conceptual models. In this paper, we report on a pilot study with college-level biology students (N=15) using VERA for modeling ecological phenomena. The 2-hour pre- and post-test study demonstrates gains in learning of ecological content knowledge. We also found that the use of Encyclopedia of Life helped students create more complex models.