Fig. 5.
Dissimilar trends in A. numerical abundance and B. mass of industrial and user plastics in fulmars from the Netherlands since the 1980s. Data show arithmetic averages ±SE by running 5-year averages. For industrial plastics, high arithmetic averages and large standard errors in the last three pentads were caused by 2 excessive outliers in 2010 and 2011. Additional log transformation in tests for trends reduces the impact of these outliers, see Online Supplement.