And here - shamelessly copied from the PP web site - is the source article which led to the question... I'll have to do this in several posts...it's long
When an athlete runs on the track, on roads or on firm ground, their legs create propulsive forces which accelerate their centre of mass and drive it forward. The athlete's centre of mass is decelerated during each recovery (early-stance) phase of the gait cycle, only to be accelerated forwards again as propulsive forces are produced by the stance leg. As they continue to run, centre of mass is accelerated and decelerated over and over again as it moves steadily forwards.
When the same athlete runs on a treadmill, centre of mass is static (at least in the forwards-backwards plane). There is no forward progress; instead, the running surface 'disappears' behind the athlete. In fact, the treadmill belt moves the athlete's legs and feet under and behind her centre of mass and, to preserve stability, their key task is to move the support leg back in front of the centre of mass in time for the impact with the treadmill belt. The key function of the leg muscles during treadmill running is not to produce propulsive forces but to re-position the legs in such a way as to keep the centre of mass stable.
Because of these major and fundamental differences, some experts have argued that treadmill training is unspecific to 'real running' and should be avoided by athletes who want to improve their running abilities on the ground. The neuromuscular patterns involved in treadmill running are so different from road, track or cross-country running, they argue, that improvements in economy and efficiency are impossible. Some go so far as to contend that treadmill training may actually impair running economy on regular surfaces.
On the other hand, treadmill advocates cite two key advantages of treadmill training: with treadmills, they argue, it is possible to perform outstanding 'hill' workouts by setting the treadmill inclination at challenging levels; treadmill athletes can also set training paces with great precision, enhancing their ability to carry out very specific training. They can do ample amounts of work at a goal race pace, for example, without ever having to worry about whether they are straying from the desired velocity.
Critics of treadmill training produce the counter-argument that treadmill running is not really very specific. For one thing, the weight and forces placed on the treadmill belt with each footstrike cause many belts to slow down briefly with every step(1). In addition, a discrete pace on a treadmill is not the same, in terms of overall physiological effort, as the same pace on firm ground; for example, six-minute miling pace on a treadmill usually produces lower heart and oxygen consumption rates than same-pace running on the track, cross-country course or soccer field. The lower cost of treadmill running is definitely attributable to lack of wind resistance; it may also reflect the biomechanical and kinematic differences between treadmill and 'normal' running.
