In two series of experiments with two different adjustments of the mechanical moulder, 50 hand-moulded and 50 machine-moulded loaves were baked in pairs for each of nine flours of various baking characteristics, using simple and bromate formulas. The data for loaf volume were analyzed statistically. With the first adjustment, machine moulding gave, for the entire series of flours, no significant reduction in variability of loaf volume when an experienced baker conducted the hand moulding. The machine-moulded loaves were significantly lower in loaf volume and exhibited less break and shred, greener crust characteristics, superior grain but inferior texture, and duller crumb color. The mean difference in loaf volume between the hand- and machine-moulded loaves baked from the different flours varied from 28 to 77 cc. or from 4.2 to 9.5% of the corresponding values for hand moulding, the difference increasing with increasing loaf volume. Machine moulding, therefore, reduced the differences between flours and between results for the simple and bromate formulas. Bakers varying m experience secured widely different mean loaf volumes by both hand and machine moulding and there was no evidence of an appreciable reduction in the variability between bakers due to machine moulding. With inexperienced operators, the variability of replicate bakings was slightly lower with machine moulding although greater than for an experienced operator moulding by hand. In general, the less the experience of the baker the lower the loaf volumes obtained. The first adjustment of the moulder did not properly seal doughs made from certain of the flours, and further experiments were undertaken to ascertain the best setting of the machine. Variations in the setting of the sheeting rolls were found to have a more pronounced effect on volume and texture than variations in the adjustment of the compression or former plate. The new setting selected was used in further comparative studies of hand and machine moulding. Replicate tests by an experienced baker using five flours baked by the simple and bromate formulas gave a slight but significant reduction in variability of loaf volume in favor of machine moulding, the coefficient of variability being 2.57% as compared with 2.88% for hand moulding. Replicate bakings of one flour on five consecutive days gave highly significant differences in mean loaf volume for the different days—the range being 18.4 and 27.2 cc. for hand and machine moulding respectively. Arranging the days in order of increasing mean loaf volume the two moulding methods gave a similar arrangement indicating that some factor or factors were operating which affected both the hand and machine results. Significant trends in loaf volume for the baking intervals within any one day were observed. These trends were not similar for the different days with the exception that the last loaves baked were uniformly lower in volume. Machine moulding did not reduce the day differences nor the trends within days, and until the cause of these secular variations is elucidated, they should be taken into account in the planning of baking tests. Experiments conducted by bakers of varying experience gave results which indicated that inexperience in moulding is not an important factor contributing to variability between replicates. The personal factor in both punching and moulding however contributes to the variability between bakers, the data indicating that moulding personality has the greater effect. Using one flour a range of 61.3 cc. was obtained between the mean loaf volumes for three bakers by hand moulding. When these bakers used doughs which were punched by an experienced operator the range was reduced to 40.1 and 8.7 cc. or hand and machine moulding respectively. From these data it is concluded that machine moulding may reduce but not eliminate the large differences in mean loaf volume which different operators secure in baking the same flour. With an experienced experimental baker, the variability of replicates for different flours was found to increase as the mean volume increased. These data are applied to a consideration of reasonable limits for agreement between duplicates and the number of replicate bakings necessary to differentiate between flours.