Defining the most adequate architecture of a bone substitute scaffold is a topic that has received much attention over the last 40 years. However, contradictory results exist on the effect of grain size and microporosity. Therefore, the aim of this study was to determine the effect of these two factors on the in vivo behaviour of β-tricalcium phosphate (β-TCP) scaffolds. For that purpose, β-TCP scaffolds were produced with roughly the same macropore size (≈ 150 μm), and porosity (≈ 80 %), but two levels of microporosity (low: 10 % / high: ≈ 25 %) and grain size (small: 1.3 μm /large: ≈ 3.3 μm). The sample architecture was characterised extensively using materialography, Hg porosimetry, micro-computed tomography (μCT), and nitrogen adsorption. The scaffolds were implanted for 2, 4 and 8 weeks in a cylindrical 5-wall cancellous bone defect in sheep. The histological, histomorphometrical and μCT analysis of the samples revealed that all four scaffold types were almost completely resorbed within 8 weeks and replaced by new bone. Despite the three-fold difference in microporosity and grain size, very few biological differences were observed. The only significant effect at p < 0.01 was a slightly faster resorption rate and soft tissue formation between 4 and 8 weeks of implantation when microporosity was increased. Past and present results suggest that the biological response of this particular defect is not very sensitive towards physico-chemical differences of resorbable bone graft substitutes. As bone formed not only in the macropores but also in the micropores, a closer study at the microscopic and localised effects is necessary.