2007ApJS..172..686K

The Naval Research Laboratory and the National Radio Astronomy Observatory completed implementation of a low-frequency capability on the Very Large Array at 73.8 MHz in 1998. This frequency band offers unprecedented sensitivity (∼25 mJy/beam) and resolution for low-frequency observations. The longest baselines in the VLA itself provide 25" resolution; the system has recently been extended to the nearby Pie Town antenna of the Very Long Baseline Array, which provides resolutions as high as 12". This paper reviews the hardware, the calibration, and imaging strategies of this relatively new system. Ionospheric phase fluctuations pose the major difficulty in calibrating the array, and they influence the choice of calibration strategy. Over restricted fields of view (e.g., when imaging a strong source) or at times of extremely quiescent ionospheric ``weather'' (when the ionospheric isoplanatic patch size is larger than the field of view), an angle-invariant calibration strategy can be used. In this approach a single phase correction is devised for each antenna, typically via self-calibration; this approach is similar to that used at higher frequencies. Over larger fields of view or at times of more normal ionospheric weather when the ionospheric isoplanatic patch size is smaller than the field of view, we adopt a field-based strategy in which the phase correction depends on location within the field of view. In practice, we have implemented this second calibration strategy by modeling the ionosphere above the array using Zernike polynomials. Images of 3C sources of moderate strength are provided as examples of routine, angle-invariant calibration and imaging. Flux density measurements of a subsample of these sources with previously well-determined low-frequency spectra indicate that the 74 MHz flux scale at the Very Large Array is stable to a few percent and that flux densities tied to the Baars et al. value of Cygnus A are reliable to at least 5%. We also present an example of a wide-field image, devoid of bright objects and containing hundreds of weaker sources, constructed from the field-based calibration. The paper also reviews other practical aspects of low-frequency observations, in so far as they differ from those encountered at higher frequencies, including aspects of interference excision and wide-field imaging. We close with a summary of lessons that the 74 MHz system offers as a model for new and developing low-frequency telescopes.