In this paper we present a multi-view registration method for aligning range data. We first align scans pair-wise with each other and use the pair-wise alignments as constraints that the multi-view step enforces while evenly diffusing the pair-wise registration errors. This approach is especially suitable for registering large data sets, since using constraints from pair-wise alignments does not require loading the entire data set into memory to perform the alignment. The alignment method is efficient, and it is less likely to get stuck into a local minimum than previous methods, and can be used in conjunction with any pair-wise method based on aligning overlapping surface sections.