Number 31

Produced, edited, and copyrighted by
Professor Brent S. Rushall, San Diego State University


Herbert G. Aas
Coach, Laurel Swim Club
East Hartford, Connecticut, USA

[The following letter and article are reprinted and edited with the kind permission of the Aas family. Herbert Aas does not enjoy good health these days. It is my honor to reproduce this eloquent article submitted to ASCA (USA) in 1986. It contains content and arguments that were well ahead of their time. It is extremely relevant to swimming coaching and technique theory now that lift force as an explanation for propulsion is being exposed for what Herbert Aas so clearly stated. This article should become part of competitive swimming lore. - Brent S. Rushall, Ph.D., R.Psy. editor]

December 18, 1986

Dear Editor:

Enclosed is an analyses of the so-called "lift-propulsion theory."

A theory is an idea, yet unproved, but needs to be thoroughly evaluated relative to existing knowledge. The lift-propulsion advocates have merely lifted selected concepts from fluid dynamics - those that support their contention. They have not only ignored other fluid dynamic concepts, which are essential in validating their theory, but they have ignored the fact that fluid dynamics is based on the principles of the science of dynamics. Lift-propulsion is not a theory. It is a speculation.

Dynamics is a sophisticated body of knowledge. Mass, force, time, distance, velocity, and acceleration are precisely defined. The methodology of determining the directional components of those quantities is well established. Yet, the advocates of lift-propulsion have ignored the procedures of dynamics and have relied on analogies, presumptions, and appearances.

For example, Schleihauf (1986), in Fig. 2 of his article in Swimming Research, defined the net hand movement as the difference between the point where the finger tips enter the water on the arm recovery and the point where the finger tips leave the water after the completion of the propulsive part of the stroke.

In the crawl stroke, at least for most expert swimmers, the fingertips enter the water before the arm is fully extended to its forward position. During the time between finger entry and full arm extension, the arm is moved forward by extending the arm and the arm moves further forward because of the forward velocity of the body. At the end of the propulsive part of the stroke, the arm is under the water. As the elbow is lifted above the water and flexed, the hand moves forward under the water for several inches. Obviously the points at which the fingers enter and leave the water is irrelevant in stroke analyses.

What is important is the velocity of the hands and arms through the water. The longitudinal velocity of the hand through the water is the difference between the velocity of the hand relative to the body and the velocity of the body through the water. Both the longitudinal and transverse velocities of the hand are essential in determining its resultant velocity and the resultant direction, which are essential in determining the angle-of-attack of the hand. It is the failure of the lift-propulsion advocates to correctly determine the angle-of-attack of the hand that has misled them into believing lift forces exist.

Schleihauf presented his experimental lift and drag results in Figs. 7 and 8. His results are not consistent with my experience - I suspect that he used a flat plate cut to a silhouette of a hand rather than a valid three-dimensional hand. But regardless of the questionable validity of his data, his data still shows the futility of attempting to use lift forces for propulsion.

In Fig. 8, the measured drag coefficients were greater than any of the measured lift coefficients shown in Fig.7. Therefore, a greater propulsive force can be applied with the hand in a propulsive-drag mode than can be applied in any lift mode. Furthermore, when his data are analyzed for the whole stroke, the non-propulsive drag forces are greater than the propulsive lift forces. If he had bothered to analyze his own data, he would not have advocated lift-propulsion.

On page 15 of the previously cited article, Schleihauf discusses what he considers "validation experiments." In each of the experiments there is a combination of longitudinal and transverse velocities - consequently there must be forces associated with both velocities. He neither makes an attempt to quantify the force components nor to quantify the correct angle-of-attack, but presumptuously uses data from his lift-drag experiments to reach his conclusions.

I have carefully read and analyzed many of the articles on lift-propulsion. I have substantial experience in advanced experimentation in the field of fluid dynamics. I can find no reason for pursuing the lift-propulsion speculation. Anyone who believes differently has every right to pursue the subject, but they also have an obligation to present a reasonable definitive qualitative and quantitative analysis before taking a position of advocacy.

Sincerely yours,

Herbert G. Aas


Herbert G. Aas

Most advances in swimming techniques have been discovered by swimmers. Since swimming became a competitive sport, coaches have tried cloning the techniques of elite swimmers into other swimmers. Sometimes the performances of the less accomplished swimmers improved, but often they did not.

As optical recording equipment became more available and as underwater photographic techniques have improved, coaches have been able to improve the accuracy of their kinematic description of the strokes of the best swimmers. It has become obvious that the movements of the arms and hands of expert swimmers are very complex. It is an important problem for coaches to determine which transverse movements are useful and which are merely flourishes that are the signature of an individual.

Eventually an imaginative observer, in studying underwater photographs of expert swimmers, noted the large and not easily explained transverse motions of the hands and arms. He speculated that perhaps lift forces contributed to the propulsion of skilled swimmers. I believe that Dr. James Counsilman has been given the credit for the initial speculation. Lift refers to the mechanism by which propellers produce propulsion and wings lift airplanes.

Since about 1970, Counsilman (1), Colwin (2), Schleihauf (3), Maglischo (4) and others have advocated the acceptance of the lift theory. I am unaware of any analyses of swimming techniques by an expert in fluid dynamics that concluded that lift contributes to swimming propulsion.

I do not know to what extent I should be considered part of the scientific community - I am an electrical engineer and for nearly 25 years I worked as a research and development engineer in the research department of a large (high-tech) corporation in the field of physical measurements. A substantial part of my work was in the design and application of instrumentation, data acquisition systems and data analyses systems for fluid-dynamic experimentation. On the basis of existing data, I neither believe that the lift theory has been proved or do I even believe that lift contributes significantly to swimming propulsion.

Lift advocates have relied almost exclusively on analogy for proof of their theory. For things that are very similar, analogy is sometimes logically useful. For things with some similarities but with large differences, analogies are mental crutches - if you cannot analyze it philosophize it,

What lift advocates have ignored is an important factor in fluid dynamics - the Reynolds number. Colwin (6) has incorrectly interpreted Reynolds number (he refers to it as the law of similarity, which he attributes to Reynolds) as a blanket license to analogize anything with anything that has some vague similarity.

The Reynolds number is a non-dimensional number. It is equal to the appropriate area of a device, multiplied by both the density of the fluid in which it is immersed and its velocity through the fluid, and divided by the viscosity of the fluid. Obviously if it is a valid concept, a change in one component of the number requires opposite changes in the other components to maintain the same Reynolds number. For example if the cross-sectional area of a sphere is doubled, the velocity of the fluid must be halved or the viscosity doubled to keep the Reynolds number constant.

The Reynolds number is not a device for analogizing Colwin's concept of similarity. It is, rather, a benchmark for predicting the interactions of identical shapes or configurations with a dynamic fluid medium. For example, the Reynolds number is an important experimental condition that permits the prediction of the flying characteristics of a full-scale airplane by testing a scale model of the plane in a wind tunnel.

Testing a model of a Piper Cub and using the results to predict the flying characteristics of a DC9 is obviously embracing disaster. Analogizing the lift characteristics of a human hand with an aerodynamically streamlined foil is further afield than that.

Under sprint swimming conditions the hands and arms are operating at a Reynolds number of approximately 100,000. This means that all non-streamlined geometries will have turbulent wakes and, consequently, high drags. The human hand has irregular leading and trailing edges, the trailing edge is as blunt as the leading edge and its two surfaces have varying curvatures and crevices. If a hand is moved in a transverse direction in such a way as to produce flow across the hand, there will be a turbulent wake as thick as the hand and the flow will be separated across most of both its surfaces. It must be remembered that for lift to exist, there must be a net flow of different velocities across the two surfaces in such a way that that the static pressures of the fluid exert a differential force on the device. With flow across the hand, lift in a fluid dynamic sense is meaningless.

With the hand at a positive angle of attack, there will be a longitudinal force (propulsion) caused by a transverse movement of the arm, but in the mode of an impulse turbine (for example spinning the propeller of a garden anemometer with a water stream from a hose). But because the hand is immersed in the fluid, at angles of attack large enough to produce significant propulsion, the drag forces will be greater than the propulsive forces. This means that more than half the energy used to produce propulsion will be lost in drag.

The lift-propulsion advocates have been misled by the stop-action effect of photographs. For example in the crawl stroke, during the stroke phase when the elbow is flexing following the arm entry, the position of the hand is frequently found to be at an angle of from 30 to 45 degrees to the transverse direction. They have incorrectly concluded that this is the angle of attack. extrapolating from photographic data in Maglischo (7), the transverse velocity during this phase is about three feet per second. By measuring the distance the body travels per stroke, it is found that the longitudinal velocity of the hand through the water for expert sprinters is also about three feet per second (this is the slippage of the hand through the water - the difference between the velocity of the hand relative to the body and the velocity of the body through the water). The correct angle of attack is the angle of the palm of the hand to its direction of motion through the water as determined from the resultant of the longitudinal and transverse velocities. In this case, the hand is at almost a 90-degree angle, therefore, there can be no lift!

Maglischo has made the reasonable speculation that possibly expert swimmers use a combination of lift and drag forces. However, when the angle that the palm of the hand makes with the longitudinal axis is related to the resultant movement of the hand through the water, what becomes obvious is this: The expert swimmer continually adjusts the hand to maximize the drag force (90-degree angle of attack) and to inhibit flow across the hand.

If lift advocates would study the concept of Reynolds number; if they would analyze the textbook diagrams that show the drag coefficients of various geometric shapes vs. Reynolds number, and the lift and drag characteristics of wings vs. angle of attack, several things would become obvious. The maximum coefficient of lift of a well-designed foil (the hand is a very poorly defined foil) is about 1.5 while the drag coefficient of a wing (rectangular plate) when the major surface is placed perpendicular to the flow is nearly 2. Obviously, using the hand and the arm in a propulsive drag-mode is both more effective and efficient. In logic, argumentation is a form of reasoning by which two related truths are used to establish a new truth. Like politicians, the advocates of lift-propulsion have made argumentation into mental hopscotch. They cite a multiplicity of truths, most of which are conditional truths, then reach the desired conclusion by jumping from one truth to another without establishing logical relationships to justify the conclusion.

For example, Colwin cites the lift on a rotating cylinder under flow conditions, the circulation theory of lift, the starting vortex in lift, the tip vortices from foils of finite length, and the ring vortices produced by flying insects (among other things). He then finds vortices along the path of swimmers' hands, depicted in photographs of expert swimmers, and declares that propulsion is being generated by lift mechanisms.

Most vortices are not produced by lift mechanisms. When fluid is displaced in an unbounded medium, regardless of the mechanism that produces the displacement, it must be replaced by circulatory flow (a vortex). The velocities of the hands and arms through the water are such (Reynolds number of approximately 100,000) that a turbulent wake is produced. The vortex trails that Colwin interprets as indicating lift are in reality turbulent eddies which are the start of turbulent or separated flows [drag pockets]. Therefore, there can be no lift in a fluid dynamic sense.

Colwin is oblivious of the need to find a specific mechanism to produce circulation before reaching his conclusion that the forearm produces propulsion through lift forces. But he, believing that vortices are the universal indication of lift forces, declares that the propulsion produced by the forearm is also lift and consequently, there must be circulation. This must be the slight of hand method of producing circulation since he speculates, ". . . that rapid manipulation of the hand on the wrist . . ." is instrumental. I for one, do not want the swimmers that I coach to waste any time during the stroke because circulation is developed slowly through the transmission of force from one molecule to the next through viscous forces.

The concept of lift-propulsion by the forearm would have been discarded immediately if Colwin had made the most elementary analysis. If any circulation did exist around the arm, at least 90 percent of the lift force generated by moving the arm through the water would be in the transverse direction since the lift force is perpendicular to the direction of arm movement. Therefore, there would be almost no lift-propulsion.

Several years ago (the mid-1930s), the crew coach at the University of Washington set the rowing world on its ear. He felt that an important part of his success was his change in stroke technique. Prior to that time, rowers were taught to start the stroke with a moderate force. The force was continually increased through the stroke and then finished with a hard impulsive force. The Washington coach taught his rowers to start the stroke (after the blade entry) with a hard impulsive force, then the stroke was continued with a strong continuous force, and completed with a small impulsive force.

A mathematical analysis of the quasi steady-state effects of the impulsive force is extremely complex, but a rational description is this. In rowing, the volume of water that is accelerated by the oar must be replaced by circulatory flow. Under steady flow conditions, because of a relatively high Reynolds number, the flow is completely turbulent on the back of the blade. The turbulence starts with the formation of eddies at the edges of the blade. The impulsive force produces a pressure wave (pseudo sound) that precedes the flow produced by the blade. The effect of the pressure wave is to broaden the shear layers at the edges of the blade and also of the surfaces of the volume of water being accelerated by the blade. This causes the replacement fluid to circulate from an area further behind the blade and in a less turbulent condition, which permits the oar to move a larger volume of water on both sides of the blade. This means that the stroke will be more efficient because there will be less slippage.

It is apparent that the more accomplished swimmers use this effect. It is especially obvious in the butterfly stroke, but it is also indicated by a quick flick of the wrist at the start of all the other strokes.

It is not difficult to develop strong quantitative evidence that swimmers are able to enhance the drag coefficients of the hands and forearms. If the longitudinal components of the movements of the hands and forearms, as determined from the photographic data in Maglischo (7), are used with the drag coefficients for flat plates and cylinders at a velocity of three feet per second, it is found that the power demand for an expert mature male is about 60 foot-pounds per second. The accuracy of this calculation is estimated to be plus or minus 20 percent. It is also estimated that the leg-kick does not provide more than an additional 10 percent of power. By extrapolating from Astrand and Saltin (9), it can be determined that the power demand for an expert male sprint swimmer is significantly greater than 100 foot-pounds per second. Consequently, it can be logically inferred that drag coefficients are substantially greater than under normal flow conditions.

Now, apply this thinking to swimming techniques. Part of the complex movements of the skilled swimmer are directed toward steering - compensation for the effects of pitch and yaw. The remainder of any transverse movements is directed toward simultaneously maximizing the swimmer's catch and maximizing strength through mechanical advantage. If the surfaces of the hands are kept perpendicular to the longitudinal axis of the body when the hands move from the maximum depth toward the body, flow will be induced across the hands, the shear layers will be thinned and the catch will be decreased. By tipping the palms of the hands toward the body, flow across the hands is inhibited to maximize the catch.

Most coaches with long experience have coached swimmers who are nearly identical in every characteristic, yet one swims significantly faster than the other does. The difference is some times attributed to their feel-for-the-water. I define a swimmer with a good feel-for-the-water as one who psychologically relates the forward speed of the body to the force exerted against the water, while a swimmer with a poor feel-for-the-water merely relates swimming speed to arm speed. Coordination can be improved with drill, swimming strength can be increased with proper exercise, but the hardest thing to teach is a good feel-for-the-water. Teaching a swimmer to induce flow across the hands and arms will not only reduce catch, but will reduce the swimmer's feel-for-the-water.

The lift-propulsion advocates have failed to demonstrate that expert swimmers generate flow across their hands or circulation around their arms. If the lift-propulsion theory were merely a matter of semantics as I am sure it is for swimmers with a good feel-for-the-water, this rebuttal would be unnecessary. But some less accomplished swimmers are being taught or encouraged by coaches and parents who do not understand the limitations of the lift-propulsion theory. The introduction of excessive transverse motions to strokes that do not produce propulsion but do waste energy is a travesty.


  1. Counsilman, J. E. (1971). The application of Bernoulli's Principle to human propulsion in water. In L. Lewillie & J. Clarys (Eds.), First International Symposium on Biomechanics of Swimming. Brussels, Belgium: Universite Libre de Bruxelles. (pp. 59-71)
  2. Colwin, C. (July/August, 1985). Essential fluid dynamics of swimming propulsion, ASCA Newsletter, p. 22.
  3. Schleihauf, R. Swimming skill; A review of basic theory. Journal of Swimming Research, 2(2), 11.
  4. Maglischo, E. W. (1982). Swimming faster. Mountain View, CA: Mayfield. (pp. 11-46)
  5. Colwin, op. cit., p. 26.
  6. Maglischo, op. cit., pp. 57-66.
  7. Astrand, P. O., & Saltin, B. (1961). Maximal oxygen uptake and heart rate in various types of muscular activity. Journal of Applied Physiology, 16, 977.

Return to Table of Contents for Swimming Science Bulletin.