Abstract Claims have recently been made once again for a test for freshness of fish based on a decrease in buffering capacity with increasing staleness. This decrease has been shown to be due chiefly to the bacterial reduction of trimethylamine oxide. However, under commercial conditions the results appear to be too irregular for such a test to be reliable. The buffering capacity of fish flesh in various ranges of pH and its variation with degree of spoilage have been almost completely accounted for by the contributions of proteins and water‐soluble constituents. The pH changes in fresh and spoiling fish, which depend in large measure on buffering capacity, are also too irregular to be satisfactorily correlated with freshness.