I got this to work using the include 'h' as an argument of the function approach. ACER13524 is correct that the value of 'h' is not taken until AFTER the procedures are called the way I had it set up which seemed obvious after I read that but it did not occur to me while I was grappling with the problem.  

The values of Q are like this . . . Each of the Q[index][3] are slight variations of the same surface equation.

Q[34];

   [x, y, 2 sin|34/25 x| cos|34/25 y| + 2/5 sin|17/25 x| + 2 cos|102/125 y|]
Q[34][3];
   2 sin|34/25 x| cos|34/25 y| + 2/5 sin|17/25 x| + 2 cos|102/125 y|

I created a surface that changes / evolves over time and made an animation of it.

Then I calulated geodesic paths that go across the changing surface while it changes. Each new geodesic takes one step on its own newly created surface with the new step location along the geodesic and the corresponding slope used as the initial values for the next geodesic calculation on the next surface.  I wanted to create an adaptive geodesic along a changing surface.

The surface animation with the adapting geodesic dot that moves along the evolving surface works really well but then I wanted to calulate the arc length of the path of the adapting geodesic dot and that is where I ran into the problem with the indexing while substituting in the for loop.

I appreciate the very quick and helpful feedback from both you, ACER13524 and TOMLESLIE1987

thanks