The maximum running speed of legged animals is one evident factor for
evolutionary selection---for predators and prey. Therefore, it has
been studied across the entire size range of animals, from the
smallest mites to the largest elephants, and even beyond to extinct
dinosaurs. A recent analysis of the relation between animal mass
(size) and maximum running speed showed that there seems to be an
optimal range of body masses in which the highest terrestrial running
speeds occur. However, the conclusion drawn from that
analysis---namely, that maximum speed is limited by the fatigue of
white muscle fibres in the acceleration of the body mass to some
theoretically possible maximum speed---was based on coarse reasoning
on metabolic grounds, which neglected important biomechanical factors
and basic muscle-metabolic parameters. Here, we propose a generic
biomechanical model to investigate the allometry of the maximum speed
of legged running. The model incorporates biomechanically important
concepts: the ground reaction force being counteracted by air drag,
the leg with its gearing of both a muscle into a leg length change and
the muscle into the ground reaction force, as well as the
maximum muscle contraction velocity, which includes muscle-tendon
dynamics, and the muscle inertia---with all of them scaling with body
mass. Put together, these concepts' characteristics and their
interactions provide a mechanistic explanation for the allometry of
maximum legged running speed. This accompanies the offering of an
explanation for the empirically found, overall maximum in speed: In
animals bigger than a cheetah or pronghorn, the time that any
leg-extending muscle needs to settle, starting from being isometric
at about midstance, at the concentric contraction speed required for
running at highest speeds becomes too long to be attainable within the
time period of a leg moving from midstance to lift-off. Based on our
biomechanical model we, thus, suggest considering the overall speed
maximum to indicate muscle inertia being functionally significant in
animal locomotion. Furthermore, the model renders possible insights
into biological design principles such as differences in the leg
concept between cats and spiders, and the relevance of multi-leg
(mammals: four, insects: six, spiders: eight) body designs and
emerging gaits. Moreover, we expose a completely new consideration
regarding the muscles' metabolic energy consumption, both during
acceleration to maximum speed and in steady-state locomotion.