The classical optical evidence for the low optical depths traditionally assigned to spiral disks is reexamined, suggesting that it is highly model-dependent and unconvincing. In particular, layered models with a physically thin but optically thick dust layer behave like optically thin disks. The opposite hypothesis, that such disks are optically thick is then examined in the light of modern evidence. It is found to be consistent with the near IR and IRAS observations, with the surface brightnesses, with the H I and CO column densities and with the H-alpha measurements.