Get reliable science

Diet composition of free roaming livestock and wildlife in extensive rangelands are difficult to quantify.� Recent technological advances now allow us to reconstruct plant species-specific dietary protein composition using fecal samples.� However, it has been suggested that validation of the method is needed through the comparison of known fed diets to laboratory results.� Using 5 heifers (2 years of age, non-gestating, non-lactating) in a 6-week feeding study (IACUC approved protocol # 20170208DS00258-01), we offered new diets weekly and collected fecal samples from each heifer after 7 days of intake. The six diets were: 1) C3 grass hay, 2) C4 grass hay, 3) C3 grass hay + C4 grass hay + alfalfa (equal proportions), 4) C3 grass hay + C4 grass hay + alfalfa (equal proportions) + minor component of sagebrush (Artemisia tridentata) leaves, 5) Alfalfa + minor components of crested wheatgrass (Agropyron cristatum), western wheatgrass (Pascopyrum smithii), and blue grama (Bouteloua gracilis), 6) alfalfa. Homogeneous diets were fed ad libitum but heterogeneous diets with multiple components were fed at metabolic weight based rations with minor diet components introduced five days prior to fecal collection.� Constrained ordination suggests that fed diets and laboratory quantified diets using DNA metabarcoding analyses of fecal samples were dissimilar.� Detection of minor components was variable.� When pooled by functional groups (C3 grass, C4 grass, legume), fecal DNA metabarcoding suggested diets were comprised of 82 to 87% of expected components and did not always align with expected proportions.� Laboratory measured protein contributions by major components was also variable when compared to expected protein contributions and at times exceeded or did not meet expectations. �Factors confounding quantification of diet composition are attributed to mis-identification in the field and the laboratory, and high numbers of �rare� species in diets attributed to fed hay sources that were not homogeneous.