The preliminary study sought to identify the main areas of the first year biology curriculum where students tended to encounter difficulties. These areas were seen to encompass potential themes where open-ended problems might be devised: light and dark reactions in photosynthesis, phytochrome and germination, cell mediated immunity, antibodies, the immunological armoury and gene function and gene expression in plant growth. The main experiment had three stages. In each stage, a battery of tests (cognitive and attitudinal) was used and the outcomes related to students' performance on an open-ended problem. Four new open-ended problem-solving units were developed. The battery of tests in the first stage 9with 642 students) were True-False tests, a Word Association Test, a Structural Communication Grid test and assessments of attitudes to learning (using a Perry Position Questionnaire). Several open-ended problems were devised and used but the outcomes from the cognitive tests were related to one individual problem-solving unit on the topic: forests that Need Fires. This stage attempted to gain some kind of insight into the extent to which nodes and links in long term memory contribute to success in problem solving. The results indicated that problem-solving success might be related to factual knowledge, links between nodes of knowledge and understanding. However, generic ability in biology might offer an alternative explanation. To test the latter, stage two involved the same sample of students in a second battery of tests, most which were on a parallel but unrelated topic: Evolution. The outcomes showed that, while some kind of generic biological ability was a factor, this did not seem to explain all the correlations observed in the first stage. The third stage approached from a fresh perspective. Working with 525 students, this involved the same four problem based units but a new set of tests was applied: tests of Lateral Thinking, a test of Convergency / Divergency, another Ranking test and a Self-Report Questionnaire. The results suggested that the number of accessible nodes and links in long term memory (reflecting 'brain architecture' and, perhaps, aspects of cognitive styles) do seem to be related to success in a problem-solving task. However, it is recognised that the validity of the test battery is an important issue in drawing such conclusions. This study raised important questions: is the way of string the knowledge or the order of storage a factor? Is it something relating to the physical factors such as structure of the brain (architecture) in terms of neurons and synapses, psychological factors such as structure of information in terms of nodes and links or emotional factors such as desire and willingness to store specific ideas other than others (knowledge filtration)? While it is well known that working memory is a key factor in determining problem-solving success (and this was not explored further in this study), the study suggests that the following factors may be important in leading to success with open-ended problems in biology: Factual knowledge held in long term memory; Understanding of conceptual knowledge; Generic biological ability (skill); Number of accessible nodes and links in long term memory; Effect of 'brain architecture'; and Aspects of cognitive styles (divergency, creativity or lateral thinking).