This article explores visual formation of science content in multimodal compositions (text and image) made by students through their second school year, and aims to contribute to the body of metalanguage for explicit description and interpretation of this kind of meaning-making. The social semiotic analysis examined both visual and multimodal meaning-making, and resulted in descriptions of seven content representations. These are representing content as theory (content reduced to its most general aspects), as a natural experience (as would be seen with one's own eyes), as an event (movement, process, perspectives), as art (abstract and/or sensory, vibrant colours and/or experimental shapes), as a person (anthropomorphization of content), as something attitude-evoking (explicitly revealing the author's attitude towards the content) and as cultural heritage (metaphoric meaning instead of scientific). Among analysed student compositions, these content representations occur both separately and in combinations with each other. Each content representation and its specific combinations of semiotic resources are presented in detail and explained with illustrative examples. Finally, the representations are discussed in terms of their relevance for teaching and assessing visually oriented science literacy in the early school years, and for pre- and in-service teacher training.