The sarcomeric M-band in vertebrate striated muscle is thought to maintain the uniform hexagonal lattice of the thick filaments by crosslinking neighbouring thick filaments in the middle of the A-band. It is also thought to maintain the fine axial alignment of the thick filaments within the A-band. The M-band has been implicated with fundamental signalling complexes that control sarcomeric protein turnover during development. There are two classes of M-band: 3-line M-band, which has a prominent central M1 stripe; and 2- line M-band that lacks the central M1 stripe. The aim of this study is to examine the ultrastructure of the different M-band types using electron microscopy and tomography. The 3-line M-band was examined thoroughly using electron microscopy and tomography of longitudinal and transverse sections. Profile plot analysis of longitudinal section micrographs and tomograms showed that the peaks of the M-band were found to be associated with protrusions on the thick filaments, rather than the M-bridges. Thus the M-bridges only contribute to the background density of the profile (envelope function) rather than the peaks. Analysis of transverse section tomograms unravelled new structural details of the M-band, which allowed us to model the possible pathways of M-band proteins. Projection of our model into longitudinal section view shows that the M-band appearance is a projection effect of the underlying 3D structure. Examination of the thick filament backbone structure showed that it is rotating at the bare region and the M-band, which suggested that the thick filament may have torsional flexibility. The path of titin in the A-band is thought to run along the thick filament in the crossbridge region, but it is not known whether titin is dislodged from the thick filament at the M-region or not. We find evidence (from the 3-line M-band specimen) that titin could be dislodged from the thick filament at the periphery of the M-region. We speculate that titin may exist as a doublet along the crossbridge segment and splits into single filaments after dislodging at the M-region. For the 2-line M-band study, freshly dissected bony fish heart muscle was prepared for plastic embedding. In longitudinal sections, the M-band did not show clear stripes but the average profile plot showed clear M-band peaks that lack the central M1 peak. Transverse sections of the M-band did not show clear hexagonal network of M-bridges. In addition, thick filaments orientations and the thick filament lattice were less ordered than in the 3-line M-band muscle. Modelling a thick filament from the raw tomograms shows a backbone structure similar to the 3-line M-band; a novel M-band model was produced, where the M-bridges originating at M4 (and M4’) do not cross the central M1 line but are oriented outwards towards the bare region. We hypothesise that M-protein is responsible for the crystalline appearance of the M-band and the thick filaments in 3-line M-bands. The work carried out in this PhD investigation will help our understanding of the structural role of the M-band in different muscles. It gives the most detailed view to date of the M-band in fast, slow and cardiac muscle. It will help our understanding of the structural changes in the M-band during activation that may initiate signalling events.