We investigated the flowering plant Salicylic Acid Methyl Transferase (SAMT) enzyme lineage to understand the evolution of substrate preference change. Previous studies indicated that a single amino acid replacement to the SAMT active site (H150 M) was sufficient to change ancestral enzyme substrate preference from benzoic acid to the structurally similar substrate, salicylic acid. Yet, subsequent studies have shown that the H150 M function-changing replacement did not likely occur during the historical episode of enzymatic divergence studied. Therefore, we re-investigated the origin of SA methylation preference here and additionally assessed the extent to which epistasis may act to limit mutational paths. We found that the SAMT lineage of enzymes acquired preference to methylate salicylic acid from an ancestor that preferred to methylate benzoic acid as previously reported. In contrast, we found that a different amino acid replacement, Y267Q, was sufficient to change substrate preference with others providing small positive-magnitude epistatic improvements. We show that the kinetic basis for the ancestral enzymatic change in substate preference by Y267Q appears to be due to both a reduced specificity constant, kcat/KM, for benzoic acid and an improvement in KM for salicylic acid. Therefore, this lineage of enzymes appears to have had multiple mutational paths available to achieve the same evolutionary divergence. While the reasons remain unclear for why one path was taken, and the other was not, the mutational distance between ancestral and descendant codons may be a factor.