Muscle volume is an important parameter in analyzing three-dimensional structure of muscle–tendon units. Three-dimensional ultrasound (3DUS) enables excellent quantification of muscle volume in small muscles;​ however, when a muscle’s cross sectional area is larger than the field of view of the ultrasound transducer at any point along its length, more than one sweep is necessary to reconstruct muscle anatomy. Confounding image registration errors have been reported between multiple sweeps. Here, we detail imaging phantom studies used to (1) define an acquisition protocol that reduces misalignment in 3D reconstruction caused by muscle deformation, and (2) quantify accuracy of 3DUS for measures of volume when phantoms are too large to be fully imaged via a single transducer sweep. Finally, we (3) establish the feasibility of our protocol for in vivo measures by comparing biceps brachii muscle volumes using 3DUS and magnetic resonance imaging (MRI). Phantom studies indicate operator intent to use constant pressure across multiple sweeps effectively mitigates image misalignment, yielding minimal volume error (1.70 ± 1.30%). Intentional application of different pressure between sweeps replicated discontinuity observed previously, leading to larger errors (5.30 ± 0.94%). Based on these findings, we adopted a gel bag standoff and acquired in vivo images of biceps brachii muscles using 3DUS and compared this volume to MRI. We did not observe misalignment errors and there were no significant differences between imaging modalities (-0.71 ± 5.03 %), indicating 3DUS can reliably be used to quantify muscle volume in larger muscles requiring multiple transducer sweeps.