I discuss both equilibrium and nonequilibrium methods for determining the masses of galaxy clusters, and point out the difficulties encountered when applying these methods to real optical and X—ray data. The best data sets now give quite precise determinations of the mass in the inner regions of relaxed clusters. The observed ratio of baryonic to total mass in these regions is at least three times larger than that predicted by the standard model for cosmological nucleosynthesis in a flat universe. This suggests either (i) that the universe is open, (ii) that the theory of light element production needs modification, or (iii) that some unknown process has led to a substantial concentration of baryonic material in clusters relative to the rest of the matter in the Universe. The degree of substructure in clusters is an indicator of their mode of formation. Significant substructure is evident in many clusters at both optical and X—ray wavelengths. Although its strength and prevalence are hard to quantify, it clearly causes a nmnber of biases both in cluster definition and in the measurement of cluster properties. I illustrate the problems using clusters selected from simulations of a Cold Dark Matter universe according to standard observational criteria. Such a model provides a good qualitative and quantitative description of observed cluster samples.