Michael,

Given that the specification of sand and silt is not unambiguous when
provided with textural classification and %clay, I wonder why you want to do
it.  Whatever your reasons, I think you are going to have to do this
yourself.

I see at least three options: Make a table of the various textural classes.
The simplest approach will just return a single set of %Sa, %Si, and %Cl for
each texture, such as the middle of the polygon.

A second approach will return a range of sand and silt for a texture and a
given clay content.  For it, there will be a max and min sand paired with a
max and min silt.  If you want, you could state min and max sand for the
clay content and compute the silt as the residual, since clay and the sand
extremes are known.  This will get somewhat extensive, but is as specific as
you can get and keep the table simple.

A third approach would provide either a line segment that describes the
family of possibilities or possibly just create a family of values
corresponding to the %sand, for instance, at the resolution you choose.  If
you want sand in integer %, then for a class with a range of 70-80% sand,
there would be a set of 11 triplets describing all the integer-resolution
combinations that stay within the textural class.

My personal choice for a solution, were the third option best, would be to
create an equation describing the sand %, given the textural class and clay
%, from the min and max sand in that class, and use it to describe the
possible combinations.  But you could perform this calculation for all
integers and make a look-up table if you wanted.

However, all the above is not going to provide a single texture for a given
class.  Instead, there will be a family of possibilities, with the number of
possibilities proportional to the area of the textural class in the textural
triangle.  So you are by definition accepting a great deal of possible error
in the texture when all you are given is class and clay.

Hope this helped.

Good luck,
John

==============================================================
E. John Sadler, Ph.D.
USDA-ARS                          [log in to unmask]
Coastal Plains Soil, Water,       http://www.florence.ars.usda.gov
   and Plant Research Center      843-669-5203x112 (voice)
2611 West Lucas St.               843-669-6970 (fax)
Florence, SC 29501-1241
U.S.A.