We investigated the behavior of participants tasked with com- municating in a novel environment. Participants had to use their mouse to draw graphical representations (termed squig- gles in the game) of human faces in order to communicate with fellow players. Experiment 1 investigated the effect of varying features of the input images on the resulting drawings. Ex- periment 2 introduced varying comprehension conditions that were predicted to produce differences in how features of faces would be graphically represented. In experiment 1, the fea- tures of the different faces significantly shaped the structure of the resulting squiggles. In experiment 2, the structure of the squiggles was influenced by the environment in which they were interpreted.