A key question concerning the use of visual illustrations in teaching is whether teaching should be adapted according to students’ preferred learning style (visualiser-verbaliser), whether focus should be on strategies that work well in general (multimedia learning), independent of preference, or whether it is worthwhile to combine the two to further improve learning. Upper secondary school students were given a lecture presented only verbally or with the aid of a visual illustration. Results from a learning test were analysed in relation to the students’ self rated learning style. Visouverbal presentation resulted in better learning than verbal presentation only, independently of learning style. Support was not found for the learning styles hypothesis, since there was no crossover interaction. However, students with mixed or visual learning styles performed generally better on the learning test than students with a verbal learning style. Since the use of visual illustrations seems to have a beneficial effect on learning for all students, this mode of instruction ought to be used in teaching. Rather than being a tool for teachers to adapt their teaching, learning styles diagnoses may be used in order to identify students who need to develop their study strategies towards a more visual preference.