Summary

The International Livestock Research Institute holds a collection of 53 accessions (genotypes) of Napier grass in its Forage Gene Bank in Addis Ababa. Napier grass has become a popular feed for livestock among smallholder farmers, because of its rapid growth, high nutritive value and yield, especially for dairy cattle. The collection shows considerable morphological variation, and so it was decided to evaluate a set of 'best-bet' accessions likely to be favoured by farmers. For example, strains with hairy leaves tend not to be favoured by farmers as they can be rough on their hands; also animals do not find hairy strains as palatable as less hairy strains.

A series of studies were therefore designed to gradually reduce the number of accessions to a manageable set that could receive final evaluation in a sheep-feeding study. This case study shows how a strategy for undertaking this research was developed, and then describes the different studies undertaken during the research process.

The first study took the form of a randomised block field experiment to compare a series of agronomic and morphological measurements in all 53 accessions. As this experiment was done in just one location it was then repeated in two locations to assess the extent of genotype x environment interactions. Nine of the better performing accessions were also evaluated in 10 trials across different countries in sub-Saharan Africa to compare their performance across different environments. The results of these three trials then finally led to a set of five 'best-bet' accessions to be evaluated when fed to sheep.

Results from the sheep study are included, as well as the protocol developed for the study to help students and young researchers to understand some of the important ingredients of a study protocol.

Methods of principal component analysis and cluster analysis applied to the morphological data collected in the first study are also illustrated.