Findings, implications and lessons learned

General conclusions from the Ukulinga field trial are:

• All landraces can be planted in October without additional fertiliser.
• Planting in December or January is too late to achieve good yields.
• Of the three landraces Dumbe-Dumbe is the best since it produces a smaller number of corms with a similar dry corm weight to the other two. This means that it produces larger corms which are easier to handle compared to the other landraces.

From a research strategy perspective the case study demonstrates the importance of setting out a plan at the beginning of a project. However, there are lessons to be learned.

• Plans do not necessarily have to be set in stone. If an early study in the research process can be analysed promptly this can guide the design of a subsequent study. This may be difficult to achieve under the constraints of a PhD programme (such as in this case study) but it is likely, as illustrated by question 8 under Study questions, that Study 2 could have been simplified if data analysis of Study 1 had been done more quickly.

• Efficient and timely data management and analysis is important. Formal statistical analysis can wait until later in the research process but simple descriptive methods used early on can often provide an adequate summary of an experiment that helps to lead the way forward and revise, where appropriate, earlier plans.

• Researchers must not forget the importance of setting out procedures for data collection and computer entry as part of the research strategy, and ensuring that there is adequate time a for this component of the work to be done in a timely way.