Furthermore, biological variations were observed across all experiments in this study. All germinated red spinach were obtained from Everything Green Pte Ltd, and they were grown in well-controlled conditions with sufficient lighting and water supply. This study did not pre-select seeds based on germination status. The inherent variability in germination rates across plants contributed to subsequent %Yield data variability, even though all samples were subjected to well-controlled growth conditions in the 9-pod gardening machines. Moreover, the inter-plant variability in morphological features, such as growth rate, may also have resulted in the overall observed variabilities. The red spinach samples randomized across the pods in the gardening machines all had a substantial variation in morphological features (e.g. leaf size, shoot length). For example, we observed that plants provided with compound combinations that resulted in their fast growth tended to limit the space and block light source for a slower growing neighboring plant. The observed trends suggested that a non-randomized design that places plants applied with the same compound in the same proximity may potentially reduce the effects of neighboring samples that exhibit different responses to the compounds. These aforementioned factors may have been the key drivers of the variability observed in %Yield data. As a result, the explanatory power and correlation for the response surfaces were not significant even though they did however provide insight into the changes in morphological features with respect to input compounds (Figure 9 and 12).