We applied two-dimensional numerical methods to describe the spatial arrangement of tubules in human dentin. The methods considered were two-point correlation functions, entropy-like measures, and angular distributions between nearest neighbors. The correlation functions were based on Fourier transform methods. The latter two approaches were based on stochastic geometry, and involved developing the Delaunay tessellations of the tubule patterns and their dual Voronoi diagrams. We discovered that for analyzing the distribution of tubules the geometric methods of lattice tessellations were more sensitive to structural order of the tubules than were Fourier-based schemes. Analysis of the data indicated that dentinal tubules are highly ordered in normal dentin.