© 2018 Elsevier B.V. Fourier transform infrared (FTIR) spectroscopy using attenuated total reflection (ATR) is commonly used for the examination of bone. During sample preparation bone is commonly ground, changing the particle size distribution. Although previous studies have examined changes in crystallinity caused by the intensity of grinding using FTIR, the effect of sample preparation (i.e. particle size and bone tissue type) on the FTIR data is still unknown. This study reports on the bone powder particle size effects on mid-IR spectra and within sample variation (i.e. periosteal, mesosteal, trabecular) using FTIR-ATR. Twenty-four archaeological human and faunal bone samples (5 heated and 19 unheated) of different chronological age (Neolithic to post-Medieval) and origin (Belgium, Britain, Denmark, Greece) were ground using either (1) a ball-mill grinder, or (2) an agate pestle and mortar, and split into grain fractions (>500 μm, 250–500 μm, 125–250 μm, 63–125 μm, and 20–63 μm). Bone powder particle size has a strong but predictable effect on the infrared splitting factor (IRSF), carbonate/phosphate (C/P) ratio, and amide/phosphate (Am/P) values. The absorbance and positions of the main peaks, the 2nd derivative components of the phosphate and carbonate bands, as well as the full width at half maximum (FWHM) of the 1010 cm−1phosphate peak are particle size dependent. This is likely to be because of the impact of the particle size on the short- and long-range crystal order, as well as the contact between the sample and the prism, and hence the penetration depth of the IR light. Variations can be also observed between periosteal, cortical and trabecular areas of bone. We therefore propose a standard preparation method for bone powder for FTIR-ATR analysis that significantly improves accuracy, consistency, reliability, replicability and comparability of the data, enabling systematic evaluation of bone in archaeological, anthropological, paleontological, forensic and biomedical studies.