[evlatests] More phenomenology

[Barry Clark]

>     The peculiar (but small) gain effects noted yesterday in the
> continuum data (i.e., the 4% square-wave-like gain changes on
> 47 second timescales in 16A, etc.) are *not present* in the spectral
> line data.  All gain solutions -- done on a 3.3 second average -- for
> all IFs are fluctuations perfectly consistent with VLA antennas. 

OK, that looks like a modest delay error coupled with the usual mild
imperfections in VLA delay tracking.