Abstract
Previous studies conducted with swine have reported that the mobile nylon bag technique (MNBT) does not always accurately predict in vivo nutrient digestibilities. Therefore, in this study, the MNBT was modified so that nutrient digestibilities would more closely resemble those from conventional (Con) digestibility studies obtained using the indicator method. A total of 19 feeds were tested including five cereal grains, five legumes, three high protein sources and six mixed diets. The principle changes to the MNBT included the use of a fecal collection harness which minimized the number of bags lost. In addition, previous protocols involved pooling of bags within pig while in the present experiment all bags were analyzed separately to increase the precision of the test. Finally, chemical analyses were done using the entire nylon bag plus residue rather than opening.the bags and scraping out the contents. With the exception of the barley sample (p=0.01), dry matter digestibility (DMD) coefficients obtained with the MNBT were not significantly different from those obtained with the indicator method. The linear regression equation relating the MNBT to the indicator method was Con DMD=-O.77+1.02 MNBT DMD ($r^2=0.93$: p<0.0001). There was no significant (p>0.05) difference in gross energy digestibility (GED) coefficients determined using the MNBT or the indicator method for any of the 19 feeds. The regression line equation relating the MNBT to the indicator method was Con GED=-5.68+1.06 MNBT GED ($r^2=0.94$: p<0.0001). The MNBT was less effective in predicting in vivo crude protein digestibility (CPD) than it was in predicting dry matter and energy digestibility. Differences greater than five percentage units were observed for two of the legumes, Kabuli chickpeas (p=0.02) and the extruded pea-canola seed mixture (p=0.01) as well as for three of the mixed diets including the unheated hulled barley-based diet (p=0.01), the unheated hulless-barley based diet (p=0.08) and the barley-soybean meal based diet (p=0.008). The regression equation relating the MNBT to the indicator method was Con CPD=5.75 + 0.90 MNBT CPO ($r^2=0.76$; p<0.0001). This study indicates that the modified MNBT can be used for the rapid determination of dry matter and energy digestibility in a wide variety of ingredients. For the measurement of crude protein digestibility, the technique produces results similar to conventional digestibility studies for cereal grains and high protein feeds but tends to overestimate protein digestibility for legumes and mixed diets.