2022AJ....163..135D

Full-polarization primary-beam patterns of MeerKAT antennas have been measured in the L band (856-1711MHz) by means of radio holography using celestial targets. This paper presents the observed frequency-dependent properties of these beams and guides users of this 64 antenna radio telescope that are concerned by its direction-dependent polarization effects. In this work, the effects on the primary beams due to modeling simplifications, bandwidth averaging, gravitational loading, and ambient temperature are quantified within the half-power region of the beam. A perspective is provided on the level of significance of typical use case effects. It is shown that antenna pointing is a leading cause of inaccuracy for telescope users in the presumed beam shape, introducing errors exceeding 1% in power near the half-power point of beams, owing to a telescope pointing accuracy of σ ≃ 0.'6. Disregarding these pointing errors, variability in the Stokes I beam shape relative to the array average is most commonly around 0.3% in power; however, the impact above 1500 MHz is on average triple that of the lower half of the band. This happens because the proportion of higher-order waveguide modes that are activated and propagate is sensitive to small manufacturing differences in the orthomode transducer for each receiver. Primary-beam correction verification test results for an off-axis spectral-index measurement experiment are included.