Corn Silage Analysis as Influenced by Sample Size

Standard deviations (%)

Size, g	CP	NDF	ADF	IVTD	NDFD
50	0.61	3.09	1.76	2.35	4.88
100	0.58	2.52	1.56	2.03	4.07
150	0.56	2.23	1.43	2.06	3.84
200	0.55	2.18	1.32	1.96	3.89
400	0.52	2.27	1.36	1.72	3.87
600	0.52	2.25	1.42	1.77	3.54

Abstract Text:

Sampling variability is a major concern when dealing with chopped corn forage that contains as much as 50% grain. Commercial laboratories prefer to minimize the quantity of corn forage that needs to be processed for analysis. Sample size and particle size of chopped forage will influence sampling variability. Our objective was to evaluate the effect of sample size collected for processing, on variability of results of forage analyses. Samples of chopped corn forage were collected from four field replicates of seven corn hybrids, in 8.3 m² plots at four sites in NY. The 112 fresh samples collected were mixed and subsampled to produce samples of 50, 100, 150, 200, 400 and 600 g. Subsamples were immediately ensiled in vacuum-sealed polyethylene bags for 30 d. An additional 600 g fresh subsample was immediately dried at 60 ^oC, and another 600 g subsample was ensiled for 30 d and then evaluated for particle size distribution. The pH of all ensiled samples was < 3.9, indicating proper ensiling. Particle size distribution using the Penn State particle separator was similar to the distribution recommended for corn silage. Crude protein, NDF, ADF, IVTD and NDFD values were analyzed using SAS PROC MIXED. Standard deviations were calculated for each site/species/size combination to evaluate variability due to sample size collected. Ensiled samples (600 g) were not different from fresh samples (600 g) for NDF, but were significantly higher (P < 0.05) in ADF and CP, and significantly lower (P < 0.01) in IVTD and NDFD. Crude protein, NDF, ADF, and NDFD all decreased quadratically with increased sample size, while IVTD increased quadratically with increased sample size. Standard deviations for all variables decreased quadratically with increased sample size. Based on both actual values and variability, a 400 g sample produced similar results to a 600 g sample, but smaller sample sizes produced both different and more variable values for all parameters evaluated.

Keywords: Neutral detergent fiber digestiblity, Sampling variation, vacuum-sealed mini-silos