Development of measurement techniques for use in a nutritive index for ryegrass breeding

Abstract

Lack of suitable criteria for early screening for nutritive value (NV) of pasture varieties has been indicated as a cause of slow improvement in lamb liveweight gain from new pasture varieties. A study was carried out to develop a NV index that could be used in the early stages of a ryegrass breeding program to rank ryegrass varieties for lamb liveweight gain. A series of four experiments, involving 12 perennial ryegrass varieties (Lolium perenne), differing in flowering date and morphological attributes, were carried out over two autumn seasons (1999 and 2001). Varieties were established in pure sward plots that were mown to a height of 5 cm, two weeks prior to experimental measurements. These swards were used to obtain data for selected NV components (NVCs) that were later used in a multiple linear regression with published lamb liveweight gain data (LWG) to identify appropriate components of a NV index. The same control variety (c.v. Nevis) was used in each experiment so relative values of NVCs could be compared with LWG. Turves of all varieties were dug from established swards and offered in pairs to trained sheep to determine preference, intake rate, bite weight and bite rate. Preference for a variety was determined as the intake of that variety as a proportion (%) of the total intake of both varieties on offer. Representative samples were taken from swards to determine chemical composition; nitrogen (N) and water soluble carbohydrate (WSC) (glkg DM), degradable N to degradable non-N organic matter (g DN/kg DNNOM) in vitro digestibility (IVDMD, glkg DM), force to break leaf (FfB; N/mg DM) and in vitro DM loss after six hours incubation (DML6; glkg DM). Methods for quantifying preference, leaf FfB, DML6 and DN:DNNOM were developed during the study in order to reflect plant resistance to grazing, rumen degradation rate and N to energy balance, respectively. These methods were shown to be precise, repeatable and able to detect significant differences between varieties for preference, leaf FfB, DML6 and DN:DNNOM (50.0 ± 10.8, 4.14 ± 0.17, 529 ± 23, and 3l.6 ± 2.0 respectively) The mean values of a range of NV components (WSC, N, IVDMD, DML6, leaf FfB, DN:DNNOM) were determined by REML analysis for 12 perennial ryegrass varieties for which lamb liveweight gain data was available from published reports. Liveweight gain was used as the dependant variable in multiple linear regressions using different combinations of NV variables as indices of NV. Two of the three indices were significant (P < 0.05) but the backward stepwise" selection resulted in two different indices in terms of the final NV components. It was concluded that neither index provided reliable coefficients because the multiple regression was based on a data set that was limited and the explanatory and response variates were derived from disparate sources. The techniques used for quantifying NVCs were simple to use and required less than 5 g DM sample material, which make them appropriate for screening at early stages of ryegrass selection. Leaf FfB was negatively correlated with intake rate (r = -0.67; P < 0.05) and therefore has potential as an indicator of resistance to grazing, handling time and perhaps even subsequent digestion rate. Sequential in vitro degradation rate (h⁻¹ ) of freeze-dried ground ryegrass was strongly correlated with degradation rate in sacco (r = 0.94; P < 0.05). Selection for ryegrass using dry matter disappearance following a single incubation of six hours in pepsin-cellulase (DML6) may improve rate of rumen degradation and emptying. The ratio between N and energy supply was defined by a DN:DNNOM measurement. A quadratic relationship for DN:DNNOM with lamb liveweight gain was observed, a relationship which supported results in the literature. Significant variation between 12 varieties was recorded for each of the NVCs measured. Previously reported heritability estimates for like-measurements (leaf shear strength, IVDMD and N content) indicate that a response to selection can be expected. It is likely that the NV measurements developed in this research would be suitable for a ryegrassbreeding index. However, it was recommended that the measurements provided in this study initially be used to select the best parent material and that directional selection for improvement of NV is used on subsequent generations. Validation of relative importance of the NV measurements was also recommended following animal performance trials on improved genotypes.