Background: This study deals with the application of concept mapping to the teaching and learning of a science topic with secondary school students in Germany. Purpose: The main research questions were: (1) Do different teaching approaches affect concept map structure or students' learning success? (2) Is the structure of concept maps influenced by gender? (3) Is the concept map structure a reliable indicator of students' learning success? Sample: One hundred and forty-nine high-achieving 5th-grade students from four German secondary schools participated in the study. The average age of participants was 10.5 years. Gender distribution was balanced. Students produced concept maps working in small, single-sex groups. Design and methods: There were two teaching approaches used: one based upon teacher-centred instruction and one consisting of student-centred learning. Both were followed by a concept-mapping phase. Student groups experienced either one or the other teaching approach. Concept map structures were analysed using of the method of Kinchin, Hay and Adams. We defined three different possible types of concept map structure: spokes, chains and nets. Furthermore, for assessing a student's short- and longer-term learning success, we constructed a multiple-choice knowledge test applied in a pre-, post-, retention-test design. Parametric tests, such as MANOVA, one-way ANOVA and t-tests were used to identify any differences in gender, teaching approach, number of nets per concept map and their interactions. Results: Type of teaching approach had an effect on concept map structure but not on students' longer-term learning success. Students of the teacher-centred approach produced more net structures than those students who participated in the hands-on instruction. Subsequent analyses showed in total more net structures for female groups. The interaction of gender and number of nets per concept map showed a significant effect on students' longer-term learning success. Conclusion: The study suggests that Kinchin's classification scheme for assessing concept map quality may be a good indicator of students' learning success when applied in combination with a knowledge test. (Contains 4 figures and 3 tables.)