By default, the three axes in a cube are treated in the same way, but for astronomical spectral line cubes the first two (RA-DEC) are different from the third (VEL). In this case the third axis is decoupled, and instead of sizes and euler angles, the largest velocity gradient dV/dR and a position angle between kinematic and morphological major axis is used.

**in=**- Input image file [???]
**clip=**- Use only values above clip, and below -clip, if only one value given. If two values given, exclude this range of values. Setting clip=0 should be sufficient to discard all pixels with the value 0.0. Default: all values used.
**wcs=t|f**- Use axis world coordinate system descriptors or the 0-based coordinate axis numbers. Default: true.
**radecvel=t|f**- Treat the 3rd axis like a separate velocity axis. This decouples the geometry from the velocity axis. Default: false.
**weight=t|f**- Use intensity weighted moments? The alternative is all pixels equal weight. Default: true.
**cross=t|f**- Should cross correlations be taken into account? Normally want this, since that is what gives the position angles, but some codes (e.g. miriad::clstats) do not do this, so for comparison this option is available. [Default: true]

% mkdisk - 1000 harmonic mass=1 | snaprotate - - 60,30 yz | snapgrid - - nz=64 zrange=-2:2 | ccdshape - clip=0 Npoints: 514 Size: 64 64 64 Center: -6.25e-05 0.0113125 -0.009375 [wcs] Theta Phi e_x: 1.00000 0.38733 -0.64866 0.65514 49.1 -59.2 e_y: 1.00000 0.86218 0.50653 -0.00821 90.5 30.4 e_z: 1.00000 -0.32652 0.56803 0.75546 40.9 119.9 qpole: 0.43241 -0.38733 0.64866 -0.65514 0.06310 0.86218 0.50653 -0.00821 0.00034 -0.32652 0.56803 0.75546

16-May-2013V0.1 CreatedPJT 19-May-2013V0.3 added weight= and radecvel=PJT