Suggested analysis

Analyse the data as a randomised block experiment with Wt as the response and No as a covariate. Prune the model of all unnecessary interaction terms and summarise.