Variations in natural abundance of carbon (C) and nitrogen (N) stable isotopes are widely used as tools for many aspects of scientific research. By examining variations in the ratios of heavy to light stable isotopes, information can be obtained as to what physical, chemical and biological processes may be occurring. The spatial heterogeneity of soil d15N- and d13C-values across a range of scales and under different land use have been described by a number of researchers and the natural abundances of the C and N stable isotopes in soils have been found to be correlated with many factors including hydrology, topography, land use, vegetation cover and climate. In this study the Latin square sampling +1 (LSS+1) sampling method was compared with a simple grid sampling approach for d13C and d15N measurement at the field scale. A set of 144 samples was collected and analysed for d15N and d13C from a 12?×?12 grid (in a 1?ha improved grassland field in south-west England). The dimension of each cell of the grid was approximately 11?×?6?m. The 12?×?12 grid was divided into four 6?×?6 grids and the LSS+1 sampling technique was applied to these and the main 12?×?12 grid for a comparison of sample means and variation. The LSS+1 means from the 12?×?12 grid and the four 6?×?6 grids compared well with the overall grid mean because of the low variation within the field. The LSS+1 strategy (13 samples) generated representative samples from the 12?×?12 grid, and hence would be an acceptable method for sampling similar plots for the measurement of mean isotopic composition.