Photographs as a direct representation of nature can serve as a means of probing pupils’ perceptions of ecosystems and biodiversity. In this empirical study, 98 pupils aged 10–12 years in primary schools in rural and urban western Sweden were asked to describe what they saw in a photograph of a lake and a forest, as well as what could be present but not visible. An ecological approach to visual perception was used as the theoretical framework. Individual pupils within groups exhibited a broad range of species knowledge. Although no animals were visible in the photograph, animals were imagined much more readily than plants. Forest species outnumbered aquatic ones in the answers. Pupils in rural schools named more tree and mammal species, while urban pupils more often used generic terms, such as ‘tree’ and ‘animal’. These findings raise implications for ‘what’ and ‘how’ in ecology instruction to reinforce plant awareness and species naming.