All averaging schemes discussed here are based on a layered earth model. The §+M approach is valid in the long wavelength limit, where the layer thickness is much smaller than a wavelength. This approach averages horizontal dipping layers and comes up with an homogeneous equivalent medium. It is essentially a low frequency wave averaging scheme. Dix on the other hand applies to infinite frequency and bandwidth. It is essentially a ray averaging scheme. Obviously surface seismic data lie in between those two extremes, potentially with some bias towards the low frequency end. I compare both schemes and will outline its possibilities and pitfalls.