Comparison of old permanent grassland, Lolium perenne and Lolium multiflorum swards grown for silage 2. Effects on conservation characteristics in laboratory silos

The aim was to compare the in-silo losses of an old permanent grassland (OP G) sward with Lolium perenne (LP) and Lolium multiflorum (LM) swards, in th eir third season post-reseeding, each managed for intensive silage producti on. The OPG sward was harvested on 22 May, 7 July, 1 September and 24 Octob er, with corresponding dates of 22 May, 6 July, 1 September,and 25 October for the LP sward and 12 May, 16 June, 19 July, 2 September and 24 October f or the LM sward. Representative samples of each of the 13 crops were precis ion-chopped and ensiled unwilted, without additive treatment. There were fi ve plastic pipe silos per harvest per sward, each containing 60 kg herbage. A wide range in chemical composition occurred among the 13 crops. Mean buf fering capacity was similar for the three swards, but water-soluble carbohy drate concentrations were lowest for OPG and highest for LM. Although all s ilages underwent lactic-acid-dominant fermentations, the differences in gra ss ensilability were reflected in mean silage pH values of 4.3, 4.0 and 3.8 for OPG, LP and LM, respectively, with corresponding ratios of lactic acid to acetic acid plus ethanol of 2.8, 3.8 and 4.0. The mean recovery rates o f ensiled dry matter (DM) across growths were 947, 955 and 913 g/kg ensiled for the OPG, LP and LM treatments, respectively, with corresponding losses of DM via effluent of 21, 26 and 36 g/kg ensiled. Silages from OPG swards generally deteriorated less during 5 days exposure to air than silages from either reseed. It is concluded that within the management systems imposed, even though the OPG sward was more difficult to preserve as silage than th e LM sward, the LP sward being intermediate, the recovery rates were better for the OPG and LP treatments than the LM treatment. This emphasises the i mportance of not comparing sward types on the basis of herbage DM yield onl y.