

   FFiinnddss ppaatthh ooff sstteeeeppeesstt aasscceenntt

        path(fit, r, x.center, x.scale, maximize=T, scale=0.0001, start, trace.z=F)

   AArrgguummeennttss::

        fit: Object from a fitting procedure. A predict func-
             tion for the fit must be available.

          r: Vector of increasing values for the radii of the
             hyperspheres.

   x.center: Center of the hypersheres, in the original coordi-
             nates of the data.

    x.scale: Scale of the hyperspheres, in the original coordi-
             nates of the data.

   maximize: Default is to search for the maximum.

      scale: Transformation scale of data.

      start: Starting value for the search, the default is the
             middle of the region.

    trace.z:

   DDeessccrriippttiioonn::

        The object from a fitting procedure must have a predict
        function.  Path calls the function nlminb which finds a
        local minimum of a smooth nonlinear function subject to
        bounded-constrained paramenters.

   VVaalluuee::

        Array with the components of the values of the data,
        the radius of the hypersphere of the search and the
        predicted value of the surface.

   SSeeee AAllssoo::

        optim, nlminb, predict.tps, predict.krig, predict.nnreg

   EExxaammpplleess::

        tps(BD[,1:4],BD$lnya,scale.type="range") -> fit # fitting a DNA strand
        # displacement amplification surface to various buffer compositions
        path(fit,c(.01,.05)) -> fit.path # path from center of data to
        # hyperspheres with radii .01 and .05

