The logarithmic transformation should be avoided for stabilising the variance of mould growth rate

Radial growth rate, μ (mm d− 1) was evaluated by plotting the radius of the colony, r (mm) versus time (d) for Alternaria alternata, Aspergillus flavus, Cladosporium cladosporioides, Mucor racemosus, Rhizopus oryzae and Trichoderma harzianum at different T and aw. For each of the 12 data sets, an analysis of variance of the raw growth rate data was performed. It was observed from the P-values that all square-root transformed values of μ were non-significant at the significance level α = 0.05, whereas for untransformed values of μ, three of the 12 values were significant and for logarithmically transformed μ, nine of the 12 values were significant at the significance level. Therefore the logarithmic transformation that is used for bacteria should be avoided for these moulds except for A. flavus. In contrast the square-root transformation appeared suitable for stabilising the variance of mould growth rate in all cases.