Snapshots require phase stability only for the duration of the
individual snapshot. There is therefore some debate over whether (for
example) an observation requiring 15 minutes of integration
time 
is better made as one continuous
15-minute observation or by combining the data from three separate
5-minute snapshots. A single 15-minute observation may give better
dynamic range, because ionospheric or
tropospheric phase gradients in the form of ``wedges'' may calibrate
out of a single short observation, leaving only a position shift. In
contrast, three shorter observations that are more dispersed in time
and calibrated on a long calibrator cycle may encounter different
wedges and so combine to give an image with poorer final dynamic
range, unless self-calibration (Lecture 9) can be used. There are
therefore some circumstances in which a single observation is
preferable, as well as being easier to schedule.
The advantages of combining data from several shorter snapshots are (a) greater protection against total loss of the data for a given source through equipment failures or short-term bad weather, and (b) more even sampling of the u-v plane than in a single extended snapshot. Multiple snapshots are particularly useful when observing at wavelengths of 18 cm and longer in the C and D configurations, as they allow better imaging of confusing sources that may otherwise limit the achieved dynamic range (see Confusion ). The single extended snapshot may be better for observations that must be made at low elevations, where phase ``wedges'' are more likely to arise, and if self-calibration cannot be used. This may be particularly true for observations of weak or complex low-declination sources for which the total hour-angle coverage is anyway limited by the short time that a given source is above the horizon.