Attempts to improve upon the crude protein feeding system have strived to characterize feedstuffs with respect to ruminal protein degradability. In vitro an in situ procedures fall short of this goal by not accounting for ruminal turnover. Six lactating cows, fitted with ruminal and duodenal cannulae, were utilized for in vivo determination of protein degradability by employing a double-marker system. Treatments were corn silage-based diets supplemented with 2:1 soybean meal:fish meal nitrogen (SF) or 1:2 soybean meal:fish meal nitrogen (FS) at levels of 15. 9 and 16.0 percent protein. Ruminal cobalt ethylenediaminetetraacetic acid (CoEDTA) infusions marked liquid phase (LP) digesta and Yb-soaked hay was used as a solid phase (SP) marker. Duodenal digesta was separated into SP and LP at 3000xg. Least squares means of LP flows did not differ for total, precipitable protein, or microbial nitrogen, for SF and FS. Total solid phase flow of dry matter was higher for FS (9.06 kg/d) than for SF (7.97 kg/d), although intakes did not differ. Milk composition and yield did not differ for FS and SF. Average daily gain was 0.12 kg/d for SF and 1:26 kg/d for FS. Whole-tract digestibilties were not different although rumen digestibility of dry matter was 24.9% for FS and 35.7% for SF, and duodenal recovery of N was 93.8% for SF and 107.5% for FS. Inclusion of a higher level of fish meal in the diet increased the proportion of feed proteins delivered to the small intestine, increasing weight gain while having no effect on milk production. / Master of Science