Abstract-The fluence of dust particles <10 lm in diameter was recorded by impacts on aluminum foil of the NASA Stardust spacecraft during a close flyby of comet 81P ⁄ Wild 2 in 2004. Initial interpretation of craters for impactor particle dimensions and mass was based upon laboratory experimental simulations using projectiles less than >10 lm in diameter and the resulting linear relationship of projectile to crater diameter was extrapolated to smaller sizes. We now describe a new experimental calibration program firing very small monodisperse silica projectiles (470 nm-10 lm) at approximately 6 km s )1. The results show an unexpected departure from linear relationship between 1 and 10 lm. We collated crater measurement data and, where applicable, impactor residue data for 596 craters gathered during the postmission preliminary examination phase. Using the new calibration, we recalculate the size of the particle responsible for each crater and hence reinterpret the cometary dust size distribution. We find a greater flux of small particles than previously reported. From crater morphology and residue composition of a subset of craters, the internal structure and dimensions of the fine dust particles are inferred and a ''maximumsize'' distribution for the subgrains composing aggregate particles is obtained. The size distribution of the small particles derived directly from the measured craters peaks at approximately 175 nm, but if this is corrected to allow for aggregate grains, the peak in subgrain sizes is at <100 nm.