Robert:

>Below is a demo of Gouraud Shading using the first approach...for FB3

Okay, what you describe is, in Petrology terms, is a "ternary" 
diagram. It is a graphical representation of the mixing of three 
items. As such, the line opposite a vertices is a zero percent 
contribution of that vertices. At the vertices is 100% contribution 
of that vertices. However, these diagrams are typically equal lateral 
triangles.

What you need is a ternary triangle where the triangle can be in any 
orientation and cofiguration, correct? One solution would be a simple 
trig algorithm, but that won't be fast. I wonder if breaking the 
triangle into smaller and smaller parts using PTINRGN would make it 
faster?

For example, I would try dividing each triangle into a set number of 
sub-triangles based upon a standard division of the sides (i.e., 1/4) 
and then using PTINRGN to determine true or false or which color to 
assign.

What do you think?

tedd