In science education, there is a now established focus on fostering students' meaning making through/as multimodal representations as part of their induction into the epistemic practices of the discipline. Increasingly, this agenda is aligned with an emphasis on nurturing students' creative reasoning, as researchers move to enrich theories of creativity with socio-cultural approaches. In this research, year 10 students interacted with three purpose-built multi-agent-based computational models (MABCM) to reason about key aspects of natural selection. I used the video coding software StudioCode to explore video footage (tripod-mounted cameras, web cameras, screencast recordings) of three pairs of students interacting in innovative and flexible ways with these multimodal simulations. I constructed and interpreted modified timelines of these reasoning episodes using Peirce's (1895/1998) social semiotics as framed by Magnani's eco-cognitive model (2001; 2009), which enabled a micro-ethnographic analysis of the multimodal nature of students' creative reasoning. I propose that students interacted with these digital simulations to engage in a multimodal (manipulative, visual and sentential) and creative (new ideas and their critique) process of hypothesising through/as abduction. I argue that students' creative reasoning as meaning making is powerfully interpreted through Peircean social semiotics as a fundamentally cross-modal process of: (1) integrating and moving between various modes; (2) logically executing formal and informal processes of proposition generation; (3) distributing meaning making across various agents (human and non-human); and ongoing hypothesising. I thus seek to add a logic as semiotic dimension to more linguistic accounts of multimodality and creativity in the science classroom.