An in situ procedure was used in four collaborative trials to evaluate the variation in estimates of ruminal escape protein among eight laboratories. A standard smooth bromegrass hay was established and evaluated in all four trials, and Trials 3 and 4 also included the standard hay material that was retained on a 75-microns sieve. Polyester bags (10 cm x 20 cm) containing 5 g of sample were soaked in water (39 degrees C) for 20 min and incubated ruminally for 16 h in a bag made of mesh material. Bag rinsing after in situ incubation consisted of a primary phase in which the mesh bags containing sample bags were washed with water (39 degrees C) in 19-L white plastic buckets. Repeated cycles of filling, agitating, and dumping were performed until the rinse water was clear. Secondary rinsing was performed on individual bags by rinsing the exterior with water (39 degrees C) followed by rinsing through the unsealed opening the interior residue to the bottom of the bag for 1 min (Trials 1 and 2). Secondary rinsing for Trials 3 and 4 was the same as for Trial 1, but with only enough water (39 degrees C) and time to rinse the residue to the bag bottom. Mean estimates of escape protein (EP/CP) were: Trial 1, 27.4 +/- 3.08; Trial 2, 30.8 +/- 3.21; Trial 3, 31.1 +/- 3.04 (non-sieved); and Trial 4, 27.7 +/- 3.06 (non-sieved). Reasonable control of error was accomplished by technician training, including the use of a videotape to clarify the procedure.(ABSTRACT TRUNCATED AT 250 WORDS)