Summary

This case study uses a multi-level design for an experiment carried out in the laboratory to determine an optimum aqueous concentration of sucrose that supports optimal germination of pollen grain from flower buds of the leguminous Sesbania sesban (L.) Merr tree species. The sucrose concentration that supports optimum pollen grain germination may vary depending on the water status of the pollen.

The purpose of this case study is to describe an experiment designed to compare pollen germination rates across a range of sucrose concentrations, and to demonstrate how analysis of variance incorporating polynomial regression analysis is applied. The optimum sucrose concentration determined from this experiment was then applied in a second experiment to determine the correlation between rates of pollen germination and viability, which was expected to be high at a high pollen germination.

Forty eight slides from six flower buds and eight sucrose concentrations were prepared in the first experiment for microscopic examination of pollen germination rates counted 15 times for each slide, each count counted in a different field of view. It is possible that alternative choices of bud and field of view replications may have been preferable in this experiment to improve the precision of germination rate comparisons across the different sucrose concentrations. The case study discusses methods of sample size estimation and uses a number of the questions at the end of the case study to investigate possible improvements in experimental design.

The case study also uses the stacking features of GenStat in order to reorganise the data for statistical analysis