A 2.2 μm speckle imaging survey of 167 bright (K < 8.5 mag) Hyades members reveals a total of 33 binaries with separations spanning 0044 to 134 and magnitude differences as large as 5.5 mag. Of these binaries, 9 are new detections and an additional 20 are now spatially resolved spectroscopic binaries, providing a sample from which dynamical masses and distances can be obtained. The closest three systems, marginally resolved at Palomar Observatory, were reobserved with the 10 m Keck Telescope in order to determine accurate binary star parameters. Combining the results of this survey with previous radial velocity, optical speckle, and direct-imaging Hyades surveys, the detected multiplicity of the sample is 98 singles, 59 binaries, and 10 triples.

A statistical analysis of this sample investigates a variety of multiple star formation and evolution theories. Over the binary separation range 01–107 (5–50 AU), the sensitivity to companion stars is relatively uniform, with ΔK_lim = 4 mag, equivalent to a mass ratio q_min = 0.23. Accounting for the inability to detect high flux ratio binaries results in an implied companion star fraction (CSF) of 0.30 ± 0.06 in this separation range. The Hyades CSF is intermediate between the values derived from observations of T Tauri stars (CSF_TTS = 0.40 ± 0.08) and solar neighborhood G dwarfs (CSF_SN = 0.14 ± 0.03). This result allows for an evolution of the CSF from an initially high value for the pre-main sequence to that found for main-sequence stars.

Within the Hyades, the CSF and the mass ratio distribution provide observational tests of binary formation mechanisms. The CSF is independent of the radial distance from the cluster center and the primary star mass. The distribution of mass ratios is best fitted by a power law q^-1.3±0.3 and shows no dependence on the primary mass, binary separation, or radial distance from the cluster center. Overall, the Hyades data are consistent with scale-free fragmentation, but inconsistent with capture and disk-assisted capture in small clusters. Without testable predictions, scale-dependent fragmentation and disk fragmentation cannot be assessed with the Hyades data.