Evaluation of a stratified continuous rumen incubation system

We present a continuous culture system based on one developed by [Teather, R.M., Sauer, F.D., 1988. A naturally compartmented rumen simulation system for the continuous culture of rumen bacteria and protozoa. J. Dairy Sci. 71, 666–673]. In which programmable stirring speeds allow a raft mat to form similar to that found in the rumen to produce a de facto dual flow system with different liquid and solid turnover rates that helps to maintain the protozoal population. Novel features include a non-clogging outflow, and a computer controlled, compact feeding system able to deliver ground substrates or pellets. The system comprises 12 thermostat controlled fermenter vessels with a working volume of 1 L each. The fermenters were primed with rumen fluid from two cows and continuously provided with the same diet as the donor animals (15 g/day for 3 days then 10 g/day) and artificial saliva (1 L/day). The concentration of short chain fatty acids (SCFA), microbial ribosomal RNA (rRNA) concentrations and total enzymatic activities were all lower in the fermenter than in the donor animals, but the proportions of SCFA remained the same. There was no difference in specific cellulase activity, but specific xylanase activity and total protozoal counts were significantly higher in the fermenters. The proportions of Bacteria, and Eukaryota (i.e., protozoa and fungi) were similar in the fermenters compared to the donor animal, but the proportion of the Archaea (methanogens) was increased. Total rRNA from cellulolytic organisms (Chytridiomycetes, Fibrobacter sp. Ruminococcus albus and R. flavefaciens) was similar to that in the donor animals. However, the composition of the cellulolytic community changed with the proportion of R. albus and the Chytridiomycetes increasing while the proportion of R. flavefaciens decreased, and the genus Fibrobacter was unchanged. Despite these differences, this in vitro system has potential to study effects of feed components on fermentation parameters. Whether the system will prove useful for modelling changes in the composition of the rumen microbial community needs to be evaluated further.