Some SA rivers and creeks have white sand. See Bleher's Discus Vol. 1
Mark
I would seriously question if any sand is actually white, as is the white pool filter sand. Photos with differing lighting can not always be taken as accurately dcepicting colour. The sands in most of the Amazon basin is pretty close to the Quikrete Play Sand, a mix of black/tan/brown/white. Given the source of the sands in the basin, white would seem impossible.
The other thing that must be remembered is that fish do not normally live in open water in sunlight. The forest canopy prevents direct sunlight from reaching the watercourses--there are obviously exceptions, portions of the Rio Guapore for instance--but not surprisingly fish do not inhabit these open areas as they are too exposed. The dorsal colour on all fish is dark, so that when viewed from above by a predator they will blend into the substrate which obviously cannot be white for this strategy to be successful.
Then there is the fact that most streams and creeks are so littered with sunken branches, roots and leaves that the actual substrate is never seen. In many cases it is mud. The basic tan colouration on so many species of Corydoras means the fish if motionless canot be seen from above. The black dorsio-lateral stripe is there to blend in with the branches on the sand. Even the orange post-orbital fleck on spcies like
Corydoras adolphoi,
C. duplicareus, etc, suggests leaves when one view the fish in the stream. Since changing over to play sand I have been aware how well these species, including C. panda, C. metae, and the spotted ones blend in. I am not sure I would have found this believable had I not witnessed it directly.
Byron.