The mechanical properties of the living, isolated anterior byssus retractor muscle ofMytilus(ABRM) have been studied under isometric conditions, and during length changes at constant speed. Stimulation of the muscle produces two distinct types of response, characterized by their tension decay rates. The rate is high in a phasic response produced by repetitive stimulation, and low in a tonic response produced by direct current or acetylcholine (ACh) stimulation. After cessation of contractile activity (i.e. the ability to shorten and develop tension actively), in both phasic and tonic responses the decay of tension becomes exponential. The time constant of tension decay ranges from 1 to 7 s for the phasic response, and from 5 to 100 min for the tonic response. Application of 5 hydroxytryptamine (5HT) to a tonically contracted muscle converts the tonic tension decay rate to the phasic tension decay rate. The resistance of the muscle to extension is greater during tonic than during phasic stimulation, but much the same during and after tonic stimulation. This is interpreted to indicate that the system which is responsible for tonic contraction is in action during tonic stimulation. Nevertheless, there is no difference in either the shortening speed at zero load, or the rate of rise of tension (after a release) during tonic and phasic stimulation; and this is also true for the muscle’s undamped series compliance. In both phasic and tonic contractions, tension is developed actively from 0*2 to 1*5l0, wherel0is defined as the shortest length of the muscle at which resting tension can be detected. Maximum tension is developed nearl0and decreases on either side of this optimum. The shape of the isometric tension-length curves is similar for phasic and tonic (ACh) contractions, but about 20% more tension is usually developed in the latter case. When the unstimulated muscle is slowly extended abovel0two types of tension are produced: (1) true resting tension, probably due to inert elastic material in parallel with the contractile apparatus, and (2) a variable amount of tension identical to passive tension (i.e. the tension remaining after stimulation when the muscle’s ability to shorten and develop tension actively is over). At any particular muscle length, the sum of passive and active tension is nearly the same, although there may be large variations in the amount of passive tension present. In a twitch, as in a tonic response, tension decays slowly, and greatly outlasts the muscle’s ability to shorten and develop tension actively. When the muscle is stimulated with single shocks spaced 10 to 30 s apart, a high level of tension can be developed, and maintained for periods up to 20 h. Such an intermittent activation mechanism may operate vivo during tonic contraction of the smooth retractor and adductor muscles of lamellibranch molluscs, thus enabling tension to be maintained very economically. Tonic tension is, in fact, often maintained in the isolated (stretched abovel0) by `spontaneous ’ contractions occurring about once every 10 s. Each individual contraction, which resembles a twitch, is accompanied by one or more action potentials similar to those observed in twitches, or at the onset of repetitive stimulation. It is suggested that tension in theABRMis developed and maintained by a system similar to that in vertebrate striated muscle, theABRMsystem being specialized in that under certain conditions tension decays extremely slowly. A hypothesis in terms of a sliding filament contractile mechanism is put forward which postulates: (1) that contraction in theABRMis due to linkages formed between two kinds of filaments, and (2) that the breaking rate of these linkages (and thus the rate of tension decay) is governed by the concentration of a relaxant (probably 5HT), the single process of breaking of linkages being reflected by the exponential phase of tension decay. The results are interpreted on this hypothesis, and discussed with reference to another hypothesis (Johnson, Kahn & Szent-Györgyi 1959; Rüegg 1959, 1961a), which holds that in addition to a contractile (actomyosin) system, muscles like theABRMhave a second (paramyosin) system which becomes rigid during tonic contraction, and thus maintains the tension developed by the contractile system.