Looks good.  Thanks again.

The problem is that this plot contains values that don't satisfly certain inequalities.  In particular, lsf*l1/(l1-lsf)<C where C is a constant (3.6 in my particular case), and lsf<l1 (although the latter inequality can be accounted for by defining the range of lsf=0..l1, as you showed previously).

Oops, theta[m]=Pi/2.

Oops, theta[m]=Pi/2.

Okay, here's the code:

>n[o]:=sqrt(2.56431+0.012337/(lambda^2-0.013103)-0.019075*lambda^2);

>n[ep]:=sqrt(2.38651+0.010664/(lambda^2-0.012878)-0.012813*lambda^2);

>n[e]:=((cos(theta[m])^2)/(n[o]^2)+(sin(theta[m])^2)/(n[ep]^2))^(-1/2);

>l2:=(l1*lsf)/(l1-lsf)

>B:=subs(lambda=l1, n[e])/l1 + subs(lambda=l2, n[o])/l2 = subs(lambda=lsf, n[e])/lsf;

>C:=subs(lambda=l2, n[e])/l2 + subs(lambda=l1, n[o])/l1 = subs(lambda=lsf, n[e])/lsf;

B and C are the functions, I'm plotting l1 against lsf.

Hmmm... No dice.  Implicitplot won't evaluate it.

Hmmm... No dice.  Implicitplot won't evaluate it.

What if I want to impose a more general constraint?  In particular, I want to impose the contraint that

x*y/(x-y) < C

where C is a constant.

It worked! Thanks a lot!