As you have superimpose trial on existing orchard. Under such circumstances we have to demarcate each unit and block where need to maintain uniformity first.

After applying treatment if you want independent result you hate test it independently. Otherwise refer the book Gomez and Gomez.

Thanks Vaibhav. I will refer to it surely.

Dear Winnie Nanjego, Coffee is grown on slopes. I suggest you set the experimental procedures using a randomized complete block design. The treatments are randomized down the slope. Use Three row plot to avoid border effects and take your data only on the center row. Use sufficient plants to get reliable data. do not analyze the end plants in the center data row. I would suggest 5 replicates to reduce the experimental error and to get a more precise estimate of it. The experiments are to set up on contour. Think about the best parameters to measure to determine your hypothesis and have your hypothesis reviewed and consult a field design statistician.

The type of experimental design to use depends on the following factors; known source of variation, number of replications, number of treatments, etc. RCBD can be used if there is only one known source of variation (wind intensity) but blocking should be perpendicular to the direction of the wind. The data can be subjected to one-way analysis of variance (ANOVA) in randomized complete block design.

The orientation of orchard also affect. light distribution is uniform to all plants. Sometimes if few plant where, southern face and others northern side. Even, if it is sloppy on sloppy land it need to be take in account that nutrient deposited at lower part.