How do I tell maxima not to evaluate diff with respect to certain variables?

Here's the problem - I want to find the Laplacian in polar coordinates,
knowing it in cartesian.  So I set up the following:

(C1) DEPENDS([U],[rho,theta,phi],[rho,theta,phi],[x,y,z])$

(C2) diff(U,x,2)+diff(U,y,2)+diff(U,z,2);


Part of what I got is:

	       2	        2		   2
     dtheta  dtheta   d U     drho     d U	 dphi	  d U
(D2) ------ (------ ------- + ---- ----------- + ---- -----------)
       dz      dz	  2    dz  drho dtheta	  dz  dphi dtheta
		    dtheta

		     2		      2			 2
   dtheta  dtheta   d U	    drho     d U       dphi     d U
 + ------ (------ ------- + ---- ----------- + ---- -----------)
     dy	     dy	        2    dy	 drho dtheta    dy  dphi dtheta
		  dtheta


Now I want to define rho, theta and phi and evalutate this thing, but
when I do that it thinks I want to take some bizarre derivative of U
and spits back an error.  Is there any way to define theta and friends,
evaluate the expression above, and not have it try to do the U
derivatives?  ev seems to be nonspecific, and I can't come up with any
other bright ideas.  I could of course define thetaprime and use subst,
but that is a long slow process.  Is there some way to handle this kind
of problem?

CY

__________________________________________________
Do You Yahoo!?
NEW from Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month.
http://geocities.yahoo.com/ps/info1