Effects of wilted grass silages varying in fermentation quality on rumen fermentation of dairy cows

Four wilted grass silages (dry matter 280 g kg-1) varying in fermentation type were fed to rumen fi stulated dairy cows to study the effect of silages on rumen fermentation. All additive treatments improved silage quality compared to untreated silage. Formic acid treated, restrictively fermented silage resulted in increased proportion of acetate, while the silages treated with Lactobacillus plantarum increased proportion of propionate of rumen volatile fatty acids.


INTRODUCTION
Inoculation using Lactobacillus plantarum VTT E-78076 (E76) has resulted in effi cient lactic acid fermentation, low pH and small proportion of NH 4 -N in wilted grass silage.However, E76 silage has been prone to aerobic deterioration.In laboratory scale silos aerobic stability has been enhanced by potassium sorbate (E202) (Saarisalo et al., 2003a).With unwilted silages an additive affects the type of silage fermentation and consequently the balance of absorbed nutrients and effi ciency of animal production (Miettinen, 1997;Huhtanen, 1998).Wilting as such restricts silage fermentation.This experiment was conducted to study the effects of silage additives used for wilted silages on rumen fermentation.
In the physiological study four rumen fi stulated dairy cows were used in a balanced 4 × 4 Latin square design with periods of 14 days.Silage was given ad libitum and concentrate 10 kg d -1 in four portions (barley 324.5, oats 324.5, sugar beet pulp 150, rapeseed meal 170 and minerals 31 g kg -1 ; crude protein 166, starch 347 and NDF 209 g kg -1 DM).Intake and milk production data from the last six days of each period were used.Rumen fermentation samples were collected on day 13 every 1.5 h from 6.00 to 18.00.Silage, milk and rumen fl uid samples were analysed according to the standard procedures of the Animal Nutrition laboratory of MTT.
Intake and milk production data and the daily means of rumen fermentation parameters were tested using the SAS GLM procedure using a model including animal, period and diet.The sums of squares for diet effects were further separated by using orthogonal contrasts into single degree of freedom comparisons: 1. NO vs additives, 2. FA vs E76 and E76S, 3. E76 vs E76S.The rumen fermentation data were tested also with MIXED procedure with a model including sampling time, time × period and time × diet interactions.

RESULTS AND DISCUSSION
In vitro organic matter digestibility of silages was 760 g kg -1 DM.Fermentation quality of the silages was good except for NO with increased amount of NH 4 -N (Table 1).However, there was a clear difference in the type of fermentation.Proportion of lactic acid of total acids was 0.81, 0.77, 0.91 and 0.91 in NO, FA, E76 and E76S, respectively.Fermentation was most extensive in NO resulting in least WSC (27.4 g kg -1 DM) and most acetic acid, butyric acid and ethanol (17.5, 1.0 and 11.0 g kg -1 DM).FA restricted fermentation resulting in highest amount of WSC (94.2 g kg -1 DM).Silage pH was 0.3 units lower with the inoculants than with NO or FA.Based on small proportions on NH 4 -N in E76 and E76S, lactic acid production was fast in the beginning of the ensiling process as has been observed in laboratory scale experiments with E76 (Saarisalo et al., 2003a).Similar fermentation quality of E76 and E76S indicates that potassium sorbate had no effect on the activity of inoculant.
On average silage intake was 12.2 kg DM d -1 and there were no differences between the silages.Extensive secondary fermentation and formation of NH 4 -N generally decrease silage intake (Huhtanen, 1998).Wilting as such restricts fermentation and therefore differences between additives are likely to be smaller with wilted silages.Energy corrected milk yield was smaller with NO than with treated silages (27.4 vs 28.3).However, the difference was not statistically signifi cant (P=0.13).Numerically milk fat, protein and lactose concentrations increased with FA compared with inoculated silages (P>0.05).Likewise in the production trial with the same silages FA tended to increase fat and protein concentrations (P<0.10;Saarisalo et al., 2003b).

EFFECT OF SILAGE FERMENTATION ON RUMEN FERMENTATION
Despite substantial difference in silage pH, rumen pH (on average 6.31) was not affected by diet, probably due to buffering capacity of saliva.Time x diet interaction was not signifi cant for the rumen parameters except for the proportion of butyrate.Diurnal variation was very small with inoculated silages while with NO and FA butyrate proportion was low before feeding but rapidly increased above the levels of E76 and E76S after feeding.
Proportion of acetate was smaller with inoculants than with FA (627 vs 645 mmol mol -1 ; P>0.01) while the opposite was observed for propionate (205 vs 180 mmol mol -1 ; P>0.01).Probably lactic acid of silage was transformed into propionate in rumen (Jaakkola and Huhtanen, 1992).Acetate and butyrate absorbed from the rumen are used for milk fat synthesis.Propionate is needed for gluconeogenesis and synthesis of milk lactose.If there is not enough propionate available, amino acids are used for gluconeogenesis (Huhtanen, 1998).

CONCLUSIONS
All the treatments improved fermentation quality of silage compared with the untreated.Potassium sorbate had no effect on function of inoculant.
Rumen fermentation was affected by the type of silage, which explains results of production trials.Inoculated silages with increased amounts of lactic acid increase proportion of propionate needed for gluconeogenesis.Restrictively fermented silage high in WSC produced more precursors for milk fat synthesis.

Table 1 .
Fermentation quality of silages, feed intake, milk production and rumen fermentation of dairy cows