Finally, the refining process increased the rate of digestion of carbohydrates, and so the onrush of blood sugar on the pancreas, which would explain diabetes. “Assume that what strains the pancreas is what strains any other piece of apparatus,” wrote Cleave and Campbell, “not so much the total amount of work it is called upon to do, but the rate at which it is called upon to do it. In the case of eating potatoes, for example, the conversion of the starch into sugar, and the absorption of this sugar into the blood-stream, is a slower and gentler process than the violent one that follows the eating of [any] mass of concentrated sugar.”

The link between refined carbohydrates and disease had been obscured over the years, Cleave and Campbell explained, by the “insufficient appreciation of the distinction” between carbohydrate foods in their natural state and the unnatural refined carbohydrates—treating sugar and white flour as equivalent to raw fruit, vegetables, and wholemeal flour. When researchers looked at trends between diet and disease, as Himsworth and Joslin had done with diabetes and Keys and a later generation of researchers would do with heart disease and even cancer, they would measure only fat, protein, and total carbohydrate consumption and fail to account for any potential effect of refined carbohydrates. Occasionally, they might include sugar consumption in their analyses, but they would rarely make a distinction between wholemeal bread and white flour, between brown rice and white. In most cases, cereal grains, tubers, vegetables, and fruits, and white sugar, flour, rice, and beer, were all included under the single category of carbohydrate. “While the consumption of all carbohydrates may not be moving appreciably with the rise or fall in the incidence of a condition,” Cleave and Campbell explained, “the consumption of the refined carbohydrates may be moving decisively.”

Cleave first made this point in 1956, when he published his hypothesis in an article that also contested Joslin’s belief that the increased incidence of diabetes in the twentieth century was unrelated to sugar consumption. Had Joslin or Himsworth charted sugar consumption separately from that of all carbohydrates, Cleave wrote, “what was the opposite of a relationship between diabetes mortality and carbohydrate consumption would become a very close relationship.”*33 (See chart on following page.)