It is made plausible for turbulent boundary layers that surface, dipole sound is in general very much less than volume, quadrupole sound. This is so for two reasons. Firstly, the relative volume sound is increased by several orders of magnitude because of shear-flow enhancement within the boundary layer. Secondly, through the use of an image argument similar to that used in the discussion of flat layers, it is indicated that the dipole sound is reduced by a factor of (L/R)2, where L is the boundary layer thickness and R is the surface radius of curvature. A discussion of irregular, surface bumps shows that such imperfections may be the source of significantly large dipole sound. A calculation is given of low-Mach-number surface and volume sound from a rotating cylinder.