HICKEY ET AL tain 667 mg tartaric acid per 100 g (11), a compound which is thought to be utilized by bacteria in the intestinal tract of man (13). T h e ripe pulp of the banana contains .50% lignin, .21% cellulose and .12°7o hemicellulose (14). Lignin is not broken down by intestinal bacteria; its presence in foods apparently prevents or retards the decomposition of both cellulose and hemicellulose(15). In spite of this, ripe bananas have been reported to be mildly laxative (16) and to increase h u m a n ileal excreta (17). This effect has been attributed to the free pectins and bound tannins found in the ripe fruit (16). T h e r e is ample evidence of a laxative effect of prunes, but the specific compound(s) responsible have not been positively identified (18), and their relationship to intestinal gas production, if any, is unknown. Most healthy people will probably experience no detected increase in intestinal gas when these fruits are eaten in customary amounts, with the possible exception of apple and prune juices. However, diets intended for patients whose gastrointestinal function is disturbed by disease or surgery might well feature apricot nectar and orange juice, with other fruits added as tolerated. Some clinical evidence relating to apple and grape juice would be helpful.

REFERENCES I. Weinstein Louis, et al: Diet as related to gastrointestinal function. JAMA 176:935, 1961 2. Calloway DH, Colasito D J, Mathews RD: Gases produced by human intestinal microflora. Nature (Lond) 212:1238, 1966 3. Calloway DH: Respiratory hydrogen and methane as affected by consumption of gasforming foods. Gastroenterology 51:383, 1966 4. Calloway DH, Murphy EL: The use of expired air to measure intestinal gas production. Ann

COMMENT MD Levitt, University of Minnesota, Minneapolis, Minn: For several centuries, flatologists (students of the subject of flatus) have argued about whether intestinal gas repre388

NY Acad Sci 150:82, 1968 5+ Calloway DH, Burroughs SE: Gastrointestinal response to diets containing pineapple. J ?im Diet Assoc 53:336, 1968 6. Calloway DH, Murphy EL, Bauer D: Determination of lactose intolerance by breath analysis. A m J Dig Dis 14:811, 1969 7. Calloway DH, Hickey CA, Murphy EL: Reduction of intestinal gas-forming properties of legumes by traditional and experimental foodprocessing methods. J Food Sci 36:251, 1971 8. Watt BK, Merrill A: Composition of Foods: Raw, Processed, Prepared. Washington, DC, United States Department of Agriculture, 1963 (Agriculture Handbook No. 8) 9. Hardings MG, Swarner JB, (;rooks H: Carbohydrates in foods. J Am Diet Assoc 46: 197, 1965 10. Shuman CR, Kemp RL, Coyne R, et al: Clinical use of sorbitol as a sweetening agent in diabetes mellitus. A m J Clin Nutr 4:61, 1956 11. Fernandez-flores E, Kline DA, Johnson AR: GLC determination of organic acids in fruits as their trimethytsilyl derivatives. J Assoc Official Anal Chem 53:17, 1970 12. Methods in Food Analysis. Second Edition. Edited by MA Joslyn. New York, Academic Press, 1960 13. Smith A, et al: The nutritional and metabolic significance of certain organic acids. J Nutr 13:601, 1937 14. Hummel FC, Shepherd ML, Mac':," IG: Effects of changes in food intakes upon the lignin, cellu~ lose and hemicellulose contents of diets. J Am Diet Assoc 16:199, 1940 15. Williams RD, Olmsted WH: The manner in which food controls the bulk of the feces. Ann Intern Med 10:717, 1936 16. Harris P, Poland GL: Effect of banana on laxation. J Lab Clin Med 24:580, 1939 17. Kramer PL: The making of high and lowresidue diets. Gastroenterology 47:649, 1964 18. Hubacher MH, Doernberg S: Laxatives. II Relationship between structure and potency..J Pharm Sci 54:1067, 1964

sents primarily swallowed air, intraluminal gas production or gas diffusingfrom the blood. The argument has swayed back and forth, with first one side and then the other appearing to obtain the upper hand. During the past 20 years the "air swallowers" appear to have gained supremacy; the

Digestive Diseases, Vol. 17, No. 5 (May 1972)

INTESTINAL GAS PRODUCTION recent literature suggests that about 70% of intestinal gas consists of swallowed air, while the remainder represents gases produced in the bowel. The present paper by Hickey et al clearly demonstrates that appreciable quantities of H2 are produced in the intestine following ingestion of certain foods and that much of this H2 is eliminated by way of the lungs rather than the anus. Similarly, the other two gases produced in large quantity in the bowel, CO2 and CH4, would also be rapidly absorbed and then excreted by the lungs. The idea that intestinal gas represents 70% swallowed air and 30% intraluminal production is derived from the composition of flatus, which consists of about 70% N2 (assumed to be entirely swallowed) and 30% H2, CO2 and CH4 (not present in the atmosphere and therefore produced in the bowel). However, a sizable proportion of the CO2, H2 and CH4 produced in the bowel will be absorbed and never appear in flatus. Thus, as a source of gas, intraluminal production doubtlessly far exceeds swallowed air. It might well be argued that it is not the rate of production (or swallowing) of gas that is clinically important, but rather the amount present in the intestinal tract at an}' given time. It is not commonly appreciated that the relative rate of passage of gases in flatus--ie, 70% N2, 30% CO~, H 2 and CH4, does not necessarily reflect the relative intestinal pool sizes of these gases~ For example, if H 2 and Nz were constantly infused at the same rate into the bowel, H 2 would be absorbed far more rapidly than would N2. The N2 to H2 ratio in rectal gas would therefore be much higher than the

Digestive Diseases, Vol. 17, No. 5 (May 1972)

relative pool sizes of these two gases. Alternatively, one might construct a hypothetical subject who swallows l0 ml of N 2 per hour and has a transit time of 1 hour. In the steady-state, 10 ml of N 2 would always be present in his intestinal tract. In addition, H2 might be produced in his colon; each time 10 ml of this gas accumulated, it would be expelled. Theoretically, this subject could pass unlimited quantities of H2 in his flatus, but the volume of H2 in the bowel would never exceed that of N2. What is therefore required to settle this longstanding argument concerning flatogenesis is the technically impossible measurement of the mean composition of all the gas present in the bowel at a given moment. In order to obtain this type of information, an attempt has been made to wash out the gas content of the bowel of normal subjects by means of a rapid intestinal infusion of argon (NEngl J Med 28.I: 1394, 1971). Analysis of the gas passed per rectum suggested that the bulk of the intestinal gas of some subjects was produced in the bowel lumen. In addition, it was demonstrated that appreciable quantities of N2 may diffuse into the lumen when the N2 concentration is reduced by intraluminal gas production. Thus, some of the N2 passed in flatus may actually be the indirect result of gas formation in the bowel, rather than of swallowed air. Clearly, the question of the origin of intestinal gas has not been entirely settled. It seems likely that a renewed challenge by the proponents of intraluminal production may be in the wind.

389