1. The purpose of the present study was to analyse the incidence and t
ype of coordination between breathing rhythm and leg movements during
running and to assess the effect of co-ordination on the running effic
iency, as well as to compare the results with those found during cycli
ng. 2. The experiments were carried out on thirty-four untrained volun
teers exercising at two work loads (60 and 80 % of subject's physical
work capacity 170) on a treadmill. In addition nineteen of the subject
s exercised at the same two work loads on a bicycle ergometer. The sub
jects were running at both work loads in three different modes in rand
omized order: with normal arm movements, without arm movements and wit
h breathing paced by an acoustic signal which was triggered by the leg
movement. 3. Respiratory variables, oxygen uptake and leg movements w
ere continuously recorded and evaluated on-line. The degree of co-ordi
nation was expressed as a percentage of inspirations and/or expiration
s starting in the same phase of the step or pedalling cycle. 4. The av
erage degree of co-ordination was higher during running (up to 40%) th
an during cycling (about 20 %) during both work loads. The difference
in the degree of co-ordination between running and cycling is probably
not due to the lack of arm movements during cycling since the degree
of co-ordination during running with and without arm movements was the
same. 5. The degree of co-ordination during running increased slightl
y but not significantly with increasing work load and could be increas
ed significantly by paced breathing. 6. The co-ordination between brea
thing and running rhythms occurred in three different patterns: (a) br
eathing was co-ordinated all the time with the same phase of step, (b)
co-ordination switched suddenly from one phase of step to another and
(c) co-ordination ensued alternatively once on the right and once on
the left leg movement. During cycling the pattern described in (a) occ
urred almost exclusively. 7. During running with a high degree of co-o
rdination, oxygen uptake for a given work load was slightly but signif
icantly lower than during running with weak co-ordination.