A technique has been developed for fragmenting striated muscle into its component thick and thin filaments by homogenization in a “relaxing medium”. Such preparations are very suitable for examination by the negative-staining technique. The thick filaments so obtained correspond closely in their structure to those seen in the A-bands of sectioned muscles. The thin filaments, often still attached to residual Z-line structures, also resemble closely those seen in sectioned tissue; they show the same characteristic internal structure, first seen and analysed in filaments from smooth muscles by Hanson and Lowy, as filaments of purified actin. Preparations of purified myosin, precipitated at low ionic strength, are found to contain spindle-shaped aggregates similar in appearance and dimensions to the thick filaments, and having a remarkable differentiated appearance along their length. Along part of the length of the filaments numerous projections are visible, probably corresponding to the cross-bridges seen in sectioned material. However, in a central zone, always about 0·15 to 0·2 μ in length, the projections are absent. The same is the case in the naturally occurring thick filaments. This appearance can be accounted for if the myosin molecule has a projection at one end, and if the myosin molecules in either half of the filaments are arrayed with opposite polarities. In confirmation of some recent results of others it is found that isolated myosin molecules, when examined by the shadow-casting technique, do indeed show such a structure. Observations on heavy and light meromyosin confirm this picture of the structure of the thick filaments.

The thin filaments and filaments of purified actin are both found to form the same very characteristic complex structure when allowed to combine with myosin, or with heavy meromyosin. The results obtained show that the filaments are structurally polarized, and in muscle are arranged so that all of them attached on one side of a given Z-line point in one direction, whilst those on the other are oppositely oriented.

A close similarity is found between the lattice structure seen in crystals of tropomyosin B and that formed by the interconnecting system of filaments at the Z-lines.

The functional implications of these results, particularly those concerned with the polarity of the filaments, is discussed.