Whales can dive for long periods by holding oxygen-binding myoglobin in their muscles at high concentrations. In the lineage leading to whales, six adaptive amino acid changes increase myoglobin's charge and folding stability, presumably to prevent precipitation at high concentrations. Higher charge increases the repulsion between myoglobin molecules, and increased folding stability reduces protein aggregation. Both traits could reduce the precipitation of myoglobin, so which trait did selection target? We aim to determine the target of selection by verifying computational estimates of folding stability and charge with empirical measurements. Computational estimates indicate that charge evolved in an improbable direction; the six amino acid substitutions in whales are rare in the space of possible mutations. Selection is the only process that pushes traits in a rare and improbable direction. This suggests that charge is the target of selection and that folding stability may have evolved as a byproduct of selection on charge. We have expressed both the ancestral and derived alleles in Escherichia coli, which can be isolated using His-tag purification. An assay to measure protein folding stability has been developed and optimized using isolated horse myoglobin. We have also developed a recombination method to generate the 64 intermediate alleles between the ancestor and derived alleles. The empirical estimates these methods generate will support a novel statistical method to verify targets of selection during adaptation. The potential applications are broad, and knowing which traits are targets of selection is critical.