One of the features that distinguish mammals from other groups of terrestri
al vertebrates is the structure and relative size of their epaxial muscles.
Yet we have only a superficial understanding of the role these muscles pla
y in locomotion. To address their locomotor function, we recorded the elect
rical activity of the iliocostalis, longissimus dorsi and multifidus muscle
s of trotting dogs. Activity was monitored at both lumbar and thoracic site
s. To develop and evaluate hypotheses of epaxial muscle function, we quanti
fied footfall patterns and sagittal trunk kinematics from high-speed videos
, and the magnitude and orientation of ground reaction forces from force-pl
ate recordings. All three epaxial muscles tended to exhibit a double-bursti
ng (biphasic) activity pattern, with the exception of the iliocostalis musc
le at the thoracic site (which was uniphasic). In general, a large burst of
activity in each muscle occurred during the second half of the support pha
se of the ipsilateral hindlimb, and was active for an average of 30 % of th
e locomotor cycle. A smaller burst of activity occurred during the second h
alf of the support phase of the contralateral hindlimb, and was active for
an average of 15 % of the locomotor cycle. Analysis of ground reaction forc
es and sagittal trunk kinematics led us to the hypothesis that the epaxial
muscles do not directly stabilize the trunk against the vertical and horizo
ntal components of the ground reaction force. Instead, the epaxial muscles
appear to counteract the tendency of the trunk to rebound (flex) in the sag
ittal plane during the latter half of the support phase. This hypothesis of
epaxial muscle function was supported by loading experiments performed on
the longissimus dorsi muscle in the lumbar region.