Functional properties of the diaphragm are mediated by muscle structure. Modeling of force transmission necessitates a precise knowledge of muscle fiber architecture. Because the diaphragm experiences loads both along and transverse to the long axes of its muscle fibers in vivo, the mechanism of force transmission may be more complex than in other skeletal muscles that are loaded uniaxially along the muscle fibers. Using a combination of fiber microdissections and histological and morphological methods, we determined regional muscle fiber architecture and measured the shape of the cell membrane of single fibers isolated from diaphragm muscles from 11 mongrel dogs. We found that muscle fibers were either spanning fibers (SPF), running uninterrupted between central tendon (CT) and chest wall (CW), or were non-spanning fibers (NSF) that ended within the muscle fascicle. NSF accounted for the majority of fibers in the midcostal, dorsal costal, and lateral crural regions but were only 25-41% of fibers in the sternal region. In the midcostal and dorsal costal regions, only ~1% of the NSF terminated within the fascicle at both ends; the lateral crural region contained no such fibers. We measured fiber length, tapered length, fiber diameters along fiber length, and the taper angle for 271 fibers. The lateral crural region had the longest mean length of SPF, which is equivalent to the mean muscle length, followed by the costal and sternal regions. For the midcostal and crural regions, the percentage of tapered length of NSF was 45.9 ± 5.3 and 40.6 ± 7.5, respectively. The taper angle was ~0.15° for both, and, therefore, the shear component of force was ~380 times greater than the tensile component. When the diaphragm is submaximally activated, as during normal breathing and maximal inspiratory efforts, muscle forces could be transmitted to the cell membrane and to the extracellular intramuscular connective tissue by shear linkage, presumably via structural transmembrane proteins.