On-farm factors relating to mycotoxin occurrence and other chemical compositional traits in grass silages in Ireland

Many pre- and post-mowing on-farm traits influence mycotoxin occurrence and conventional chemical compositionin grass silages. The objective of this study was to determine if observed on-farm pre- or post-mowing factors can be used to indicate mycotoxin occurrence or conventional chemical composition traits in grass silages. Over a two year period (2012-2013 and 2013-2014) 290 grass silages were sampled on 145 farms and these consisted of baled (n=115) and pit (n=175) silages. On each farm in each year, a questionnaire was completed and pre- and post-mowingfactors recorded. The preservation quality was better in pit compared to baled silages in both years and in silages made in Year 2 compared with Year 1. On-farm pre- and post-mowing factors were generally not secure indicators of silage conventional chemical characteristics. In contrast, the occurrence of enniatins and beauvericin generally increased with harvesting month (May-October) for grass silages. Visible mould on baled grass silages and a rottedtop layer with pit grass silages were associated with andrastin A, mycophenolic acid and roquefortine C occurrence.

Variation associated with sampling bale or pit silage for mycotoxins and conventional chemical characteristics

Mycotoxins are heterogeneously distributed analytes such that obtaining a representative sample of silage for analysis can be highly challenging. The objectives of this study were to: (1) identify the variability in mycotoxin concentrations, nutritional value characteristics and fermentation characteristics when 20 cores were taken from individual next-to-be-fed silage bales and five cores were taken from the next-to-be-fed section of the feed face of pit silages, (2) compare the values obtained from intensive sampling of the silage after simulated feed-out to values obtained by the above mentioned core sampling, and (3) describe the impact of reducing the number of core samples taken from the baled and pit silages. Round baled silages (n=10) or sections (2 m wide × 1 m height) of pit silages (n=10) were core sampled at twenty and five positions, respectively. After coring, baled silages were chopped, mixed and placed along a simulated feed trough. Silage in the simulated feed trough was then grab sampled (n=20) to produce an aggregate sample, and this was undertaken in triplicate. For pits, silage was mixed and placed along a simulated feed trough before being sampled, as per baled silage. Variation within or among either baled or pit silages, and for core or feed trough samples, was generally much higher for mycotoxins than for conventional chemical composition traits. Within silage (bale or pit) and among silages, variation was generally reduced when samples were collected from the feed trough, except for mycotoxins in pit silages which had higher variation in the feed trough compared with core samples. Increasing the number of core samples increased the likelihood of detecting a positive sample if present and precise estimates of the true chemical composition of the population. Representatively sampling a next-to-be-fed section or bale of silage for conventional chemical characteristics was achieved with 2-4 core samples or 1-2 feed trough samples. However, in contrast, when sampling silages for mycotoxins, collecting a representative sample using core or feed trough sampling methods sometimes required over 100 and over 20 samples, respectively.