Add the current model to the continuum model.
EXAMPLE
-------

0>setcont
Added 20 components to the continuum model.
0>

CONTEXT
-------
In observations of spectral-line sources, background continuum is
often present. If the spectral-line features of the source are of
primary interest then the constant contribution from the continuum
part of the source should be removed before mapping individual
spectral-line channels.

The usual way that this is done is to map channels that do not contain
any spectral line features. This results in a clean-component model of
the continuum. This model is then subtracted from the channels that do
have spectral-line emission, leaving only the spectral-line features
to be mapped.

While, in theory the subtraction of the continuum model could be done
either in the map plane, or in the UV plane, in general it is more
accurate to do it in the UV plane, and this is what difmap allows you
to do.

Difmap differs from other packages in the way that it does this, in
that it employs residual mapping techniques to make continuum
subtraction more flexible. In difmap the continuum model is never
subtracted directly from the observed visibilities. Instead difmap
maintains the continuum model as a part of the established model,
(which is subtracted on-the-fly during gridding). When the residual
map is restored to display the clean map, the continuum portion of the
established model is not added into the map along with the rest of the
model, thus leaving just the spectral-line portion of the map.

The advantages of this approach are:

1. The continuum model can be adjusted or deleted at any time without
   affecting the observed UV data.

2. Since the model visibilities and the observed visibilities are kept
   apart, it is still possible to use self-calibration after a
   continuum model has been installed.

3. It may be quicker given that the subtraction is only performed on
   channels that are actually mapped.

The disadvatages are:

1. It is not possible to take continuum-subtracted visibilities into
   another package, since the model is saved as a model file separate
   from the UV FITS file.

Another feature that may be seen as an advatage by some, and a
disadvantage by others, is that radplot and other visibility plotting
commands, display un-subtracted visibilities.

USAGE
-----
The 'setcont' command adds the current tentative and established
models to any existing continuum model.

For example, say one had a source with spectral-line emission only in
channels 21-29 of a 40 channel observation:

1. To determine a continuum model one would use the select command to
   select channels 1..20 and channels 30..40, since these channels
   only contain continuum. 

2. Make the best possible map from these channels, using clean or
   modelfit, combined with selfcal if the continuum is strong enough.
   All the selfcal corrections are applied across all channels, so in
   correcting the continuum channels you also correct the
   spectral-line channels.

3. Use the setcont command to add the model made in (2) to the
   continuum model.

4. To check the continuum subtraction, select a single one of the
   continuum only channels, and display the residual map with mapplot.
   If this reveals any structure then the continuum model is not
   correct, and further cleaning etc.. should be applied. Go back to
   step 1, and perhaps try selecting a better set of continuum
   channels. If you do any further cleaning, then use the setcont
   command to update the continuum model with the new components.

5. If the continuum was too weak for self-calibration to be feasible,
   but one or more of the spectral-line features is strong enough,
   select the strong channel, make a map of it and perform self-cal as
   necessary. If significant corrections were made then, go back to
   step 1, to improve on the continuum model.

6. Map each channel containing spectral-line features. Use the

    mapl cln

   command to see the clean map, uncontaminated by the continuum.

OTHER COMMANDS THAT AFFECT THE CONTINUUM MODEL
----------------------------------------------

wmodel:

 With the addition of the continuum model, difmap now has three models
 that can be addressed by the user. For instance the clrmod command
 allows you to specifically select which of these get cleared.

 After using the select command, you are advised to type:

  0>clrmod true

 This clears both the established and tentative models that were built
 up during the previous channel selection, but preserves the continuum
 model for the new channel selection.

 If you want to clear the continuum model, the optional 3rd argument
 of clrmod must be given as 'true'.

wmodel and rmodel:

 The wmodel and rmodel commands default to writing/reading the
 non-continuum parts of the established and tentative models. The
 optional second agument is required to ask it to separately handle the
 continuum part. See help wmodel and help rmodel, for details.

save and get:

 The 'save' command saves continuum models separately from the normal
 models, in a file with suffix ".cmod". The 'get' command looks for a
 file with this extension and if found, reads the file, placing the
 model directly in the continuum model.