I read with interest the recent article of Hopker et al. (5), commending the Douglas bag technique (4) as the “gold standard” approach when determining the gross mechanical efficiency during cycling. Their article provided an excellent analysis of many of the potential sources of error when using a Douglas bag, but the issue of changes of gas composition within the bag before chemical analysis would have profited from a consideration of both the physical principles involved and previous publications on this topic (3,6,7). Account must be taken of oxygen, carbon dioxide, and nitrogen. Moreover, the rates of gas exchange depend upon the type of bag used (rubberized fabric or neoprene), its age, its physical characteristics (bag volume and surface area relative to the volume of stored gas), partial pressure gradients, and environmental temperature.
In theory, two distinct physical processes could lead to a change in gas composition: 1) unrestricted molecular movement through relatively large apertures and 2) more restricted movement through pores of molecular dimensions. According to kinetic theory, the loss of gas through large apertures should be related to the square root of the partial pressure gradient, whereas losses through molecular pores should show a linear relationship to the pressure gradient. Shakespear (6) tested dirigibles that possibly had sustained rough handling, and he concluded that the first process was dominant. However, observations on rubberized Douglas bags showed a direct relationship to the driving pressure (7).
The relative rates of gain of nitrogen and losses of oxygen and carbon dioxide for a unit pressure differential were 1.0:2.4:10.0 (7). These ratios correspond closely with solubility of the three gases in rubber (1,2,6,8), suggesting that gas exchange occurs in three stages: 1) solution of gas at the proximal surface of the bag, 2) diffusion through the fabric, and 3) evaporation at the distal surface. Because the respiratory gases are less soluble in neoprene than in rubber, losses are smaller for neoprene than for rubberized fabric bags.
The effect of mass transfers upon gas concentrations within the bag depends on the bag volume, its surface area, and the initial volume and concentration of gas. If 5% CO2 is stored in a well-filled 100-L bag with a surface area of 1.41 m2, the rate of loss is about 50 mL·h−1 (7), causing the concentration of CO2 within the bag to change by about 0.05%·h−1. This is somewhat greater than the 0.015%·h−1 observed by Hopker et al. (5), but presumably, these authors were using a neoprene bag.
The end conclusion is as indicated by Hopker et al. (5). If gas is sampled within 15 min of collection, the error from bag leakage will remain within the error of most chemical analyses.
Roy J. Shephard, M.B.B.S., M.D. [Lond.],
Ph.D., D.P.E., LL.D.
University of Toronto
Brackendale, British Columbia
1. Barrer RM. Diffusion in and through Solids. Cambridge (UK): Cambridge University Press; 1951. p. 464.
2. Daynes HA. The process of diffusion through a rubber membrane. Proc R Soc Lond A Contain Pap Math Phys Character. 1920; 97: 286–307.
3. Director of Air Services. The permeability of airship fabrics. Tech Rep Advis Comm Aeronaut Lond. 1918–1919; 3: 1316–37; 1338–63.
4. Douglas CG. A method for determining the total respiratory gas exchange in man. J Physiol. 1911; 42: 17–8P.
5. Hopker JG, Jobson SA, Gregson HC, Coleman D, Passfield L. Reliability of cycling gross efficiency using the Douglas bag method. Med Sci Sports Exerc. 2012; 44 (2): 290–6.
6. Shakespear GA. A new permeability tester for balloon fabrics. Tech Rep Advis Comm Aeronaut Lond. 1916–1917; 2: 579–83.
7. Shephard RJ. A critical examination of the Douglas bag technique. J Physiol. 1955; 127: 515–24.
8. Venable CS, Fuwa T. The solubility of gases in rubber and rubber stock, and effect of solubility on permeability. J Ind Eng Chem. 1922; 14: 139–42.