Skip to main content

How Long Will The Podiatric Myths Of The Midtarsal Joint Survive?

The midtarsal joint is probably one of the least understood but most important joints of the human foot. During daily weightbearing activities, the midtarsal joint must be compliant enough to allow the foot to adapt to uneven terrain and stiff enough to allow the powerful gastrocnemius and soleus muscles to push the body efficiently forward off the plantar forefoot. The ability of the midtarsal joint to biomechanically optimize the load versus deformation characteristics of the foot is what makes this pedal joint such an important part of the human bipedal locomotor apparatus.

One of the reasons the midtarsal joint has been so poorly understood over the years is because, until recently, there has been little scientific research to shed light on its biomechanical function. In 1941, Manter was the first to describe two simultaneously occurring midtarsal joint “axes,” the oblique transverse tarsal and longitudinal transverse tarsal axes.1 Manter moved non-weightbearing cadaver feet in motion patterns that he believed were the most physiologic.

In 1953, Hicks also studied cadaver feet and found that the oblique and an-terior-posterior midtarsal joint axes were in different locations than Manter’s axes.2 Unfortunately, Manter and Hicks didn’t describe the direction or magnitude of the input force (i.e., external force applied to the foot) that produced their “joint axes.” Knowledge of the input force used is crucial to understanding how these researchers arrived at the location of their midtarsal “axes.”  

Then, in 1960, Elftman used his “eye of a connoisseur” method to inspect the joint morphology of the midtarsal joint visually and claimed there was a major and minor “axis of curvature” in both the talonavicular joint and calcaneocuboid joint.3 He proposed that when the major axes of the talonavicular joint and calcaneocuboid joint were parallel to each other with the subtalar joint pronated, there would be more midtarsal joint motion available. Unfortunately, podiatric schools around the world are still teaching this idea that alignment of visually determined midtarsal joint “axes” are what make the midtarsal joint more compliant in pronated feet, even though there is not a shred of research evidence that supports Elftman’s theory.

Thankfully, modern scientific research has provided us with more reliable data to better understand the biomechanical function of the midtarsal joint. In 1983, Van Langelaan implanted metallic beads into 10 cadaver feet with three-dimensional X-ray photogrammetry allowing precise determination of the motions of the talonavicular and calcaneocuboid joints.4 He found that rather than the configuration suggested by Manter and Hicks, both the talonavicular and calcaneocuboid joints had joint axes that constantly moved within space, independent of each other, during pronation and supination of the weightbearing foot.

In more recent research on live patients using 3D motion analysis, Nester and colleagues also refuted the notion that two simultaneously occurring oblique and longitudinal midtarsal joint axes can exist since the same bones can’t be moving in two different directions at the same time and since “axes of rotation do not determine the motion at a joint; rather the motion determines the axis.”5

Additionally, the latest 3D bone pin research in six patients during walking found the talonavicular joint has just as much motion during walking as the ankle joint, and the cuboid-fifth metatarsal joint has more motion than either the calcaneocuboid joint or subtalar joint during walking.6

With these facts in mind, why then is the podiatric profession still teaching this 60-plus-year-old concept when modern biomechanics research has long since disproven these ideas? Likewise, why is the podiatry profession teaching Elftman’s theory — that the alignment of his imaginary talonavicular and calcaneocuboid joint “axes” increase the midtarsal joint range of motion when the subtalar joint is pronated — when there is no research evidence to support it?  

It is now time that we, as a profession of foot health specialists, spend more effort keeping current on the latest biomechanical research so we can stop teaching the myths that have long been passed along from one generation of podiatrists to another. This is the best way for us to retain our status as the biomechanics specialists of the foot and ensure that our patients continue to receive the best conservative and surgical care.

Dr. Kirby is an Adjunct Associate Professor within the Department of Applied Biomechanics at the California School of Podiatric Medicine at Samuel Merritt University in Oakland, Calif. He is in private practice in Sacramento, Calif.


  1. Manter JT. Movements of the subtalar and transverse tarsal joints. Anat Rec. 1981; 80:397-410.
  2. Hicks JH. The mechanics of the foot. I. The joints. J Anat. 1953; 87(4):25-31.
  3. Elftman H. The transverse tarsal joint and its control. Clin Orthop. 1960; 16:41-44.
  4. Van Langelaan EJ. A kinematical analysis of the tarsal joints. Acta Orthop Scand. 1983; 54(Suppl 204):135-229.
  5. Nester CJ, Findlow A, Bowker P. Scientific approach to the axis of rotation of the midtarsal joint. J Am Podiatr Med Assoc. 2001; 91(2):68-73.
  6. Lundgren P, Nester C, et al. Invasive in vivo measurement of rear-, mid- and forefoot motion during walking. Gait Posture. 2008; 28(1):93-100.
Kevin A. Kirby, DPM



Dr. Kirby's brief description of the work of Manter, Hicks and Elftman in constructing the model of the midtarsal joint axes is basically correct. There are though some misconceptions about the work of these early publishers and also the work of the later researchers Kirby references. I would also like to add some history publicly to the dialogue. Some of these points about the midtarsal joint I have already covered in the chapter that I contributed to the book "Lower Extremity Biomechanics: Theory and Practice, Volume 1," (ed. Stephen F. Albert and Sarah A. Curran, Bipedmed, LLC, Denver, Colorado, 2013). During a recent dinner I had in the company of Ray Anthony, writer of the of book "Manufacture & Use of the Functional Foot Orthosis," (Karger AG, Basle, 1990), he related the following story about his meeting with Dr. Root soon after the publication of the book. Dr. Root told Mr. Anthony that he would come to regret the writing of a book because no one in the future would ever let him change his mind. Indeed, Dr. Root made the following comment to a class that he taught in 1979 at the California College of Podiatric Medicine, "I only believe 50% of what I taught 5 years ago and I think that I will only believe 50% of what I teach today 5 years from now." First, it is important that the Elftman concept of midtarsal joint motion not be confused with the axes of motion described by Manter and by Hicks. Manter and Hicks discussed axes of motion as defining the capability of the joint to move in certain pre-defined directions. It may be more correct to say that this 2 axis model determined the degrees of freedom that the midtarsal joint had to move. On the other hand, the axes of the talonavicular and the calcaneocuboid joints discussed by Elftman was a geometrical axis based purely on a frontal plane projection of the joint shapes. Such a projection shows that the talar head is an oval, which has a long axis and a short axis. He also drew the anterior facet of the calcaneus as a rounded rectangle with a long and short axis. He attributed the limitation of motion of the midtarsal joint in the pronation direction as being due to the orientation of these two geometrical axes. He also stated that the functional axis motion of the midtarsal joint was determined by the cross product of the unit vectors of the two geometrical axes. For the record, I give this personal account from my own early involvement. In 1975, I was in the employ of Dr. Milton Wille, a Professor of Mechanical Engineering at Brigham Young University, as he worked with my father, Robert L. Phillips, DPM, to develop a better material for making orthotics. To increase his knowledge of podiatric concepts in biomechanics, Dr. Wille flew to San Jose, California to meet with Dr. Root. Of great interest to Dr. Wille was Root’s adoption of the Elftman model that described why the midtarsal joint had less motion when the subtalar joint was in a supinated position and more motion when the subtalar joint was in a pronated position. At the end of their discussions, three days later, Dr. Wille informed Dr. Root that the Elftman model had no mechanical basis and challenged him to an anatomical demonstration. Dr. Wille designed the jig to do the testing and Drs. Root and Weed obtained a freshly amputated limb. When the limb became available, Dr. Wille flew to San Jose where he, Dr. Root, Dr. Weed and Dr. Orien did the dissection work and loaded the foot in a simulated weightbearing condition while the ligaments crossing the midtarsal joint were sequentially cut. The Elftman model stated that with the subtalar joint in a supinated position, the midtarsal joint would hold its integrity without the ligaments. This was often called the “osseous locking mechanism” of the midtarsal joint. Dr. Wille maintained that as the ligaments were cut, the arch would get flatter and flatter. The experiment proved that Dr. Wille was indeed correct. I do have in my possession a letter written by Dr. Root in 1976 giving the protocol for this experiment, however it is most unfortunate that the results of this experiment were never published by these authors. I discussed the experiment with Dr. Wille soon after his return from doing it and have also discussed in the years since with Dr. Root and Dr. Orien. For the historical record, the book “Normal and Abnormal Function of the Foot” in 1975 was ready for publication, however Dr. Root in his effort to be accurate, held up publication of the book in order to add a chapter on basic mechanics by Dr. Wille and also to remove the term “osseous locking” from all locations in the text based on the results of this experiment. As a result, the copyright date on the book is 1977 instead of the originally planned 1975. To jump forward a little, in 1987, I attended one of the annual Root seminars in Sacramento, California. At that conference, in a private discussion with Dr. Root, I told him that my clinical observations were that the midtarsal joint had three axes of motion, not two, and that these three axes of motion were not necessarily orthogonal to each other. I had no measurements to prove this concept as it was based only on my observations. I was indeed embarrassed when the next day, Dr. Root, in his excitement about what I had told him, publicly announced my concept that I believed there were three axes of motion. I still don’t have the equipment to prove or disprove the concept I expressed to Dr. Root. In my discussion with Dr. Root, he also relayed to me that when he and his colleagues were doing their research to try to replicate the work of Hicks, (published: Root ML, Weed JH, Sgarlato TE, Bluth DR. Axis of motion of the subtalar joint. J Am Podiatry Assoc. 1966;56(4):149-55), they agreed with Hicks on the subtalar joint axis, but that they could not replicate Hicks’ findings on the midtarsal joint having two independent axes of motion. I share these incidents to demonstrate that Root was not stuck in just trying to preserve an antiquated concept, but he was always looking for more information to increase our knowledge base. Now let’s look at some misconceptions that Dr. Kirby would have us believe. In particular, I point out a misconception that comes from the work by Nester, et al. (Nester CJ, Findlow A, Bowker P. Scientific approach to the axis of rotation of the midtarsal joint. J Am Podiatr Med Assoc. 2001; 91(2):68-73). This work purports that there are not two axes of motion of the midtarsal joint but that there is a single axis of the midtarsal joint, which moves around when one walks. Dr. Nester came to that conclusion, not by trying to move the midtarsal joint in various directions when nonweightbearing, but instead calculated the axis of motion during the gait cycle. To understand why this concept cannot identify how many axes there are in the midtarsal joint, let us look at the model of hip joint motion, which is about as close to being a true ball-in-socket joint as exists. This joint is often said to have 3 axes of motion that are orthogonal. As such, we can describe all motions of the hip joint as being the summation of a certain amount of motion around one axis plus another quantity of motion around the second axis plus a third quantity of motion around the 3rd axis. If you see the hip joint move in a certain direction (e.g. the hip abducts with external rotation while flexing), would you see motion occurring sequentially around the three cardinal axes? The answer is ‘no’. You could not calculate the three axes of motion from observation but instead, you would determine that the complex motion had occurred around a single joint axis. If you watch a person walk and determine the axis of motion of the hip joint, similar to the way that Nester did in the midtarsal joint, you would conclude that the hip had a single moving axis of motion that was not aligned with any of the three mathematical axes. In likewise similarity, the work of Nester - which indeed is commendable and should be studied – cannot find how many axes of motion there are in midtarsal joint. It only gives a summation of motion around the various mathematical axes. Nester does not negate the findings of Hicks. It is unfortunate that Nester, et al. did not break down their single moving axis to find out if it was possible to describe their findings in the mathematical terms of Hicks, or whether a third axis needs to be introduced. The two axis model has certainly been extremely useful in trying to model closed kinetic observations and trying to understand why certain pathologies occur. However, it is just that: a model. It simplifies teaching in a medical educational system that has continually de-emphasized the importance of a mathematical and mechanical knowledge base. Indeed, a great many podiatric students do not have the mathematical skills to read the research papers on midtarsal joint kinesiology. The two-axis model also dictates certain treatment paradigms that have for the most part been very successful in clinically treating symptoms. This treatment paradigm still plays an important part of Dr. Kirby’s day-to-day clinical practice, whenever he utilizes the Root technique of casting a foot for custom orthotics. One of the problems with all of the newer research that Kirby cites in his editorial is that none of it has yielded any changes in the basic treatment paradigms nor has better explained any of the common podiatric pathologies. Just as Root told me in 1987, the Hicks model is not accurate and just as he was looking in 1987 for a better model, we continue to look for a better model. The purpose of any model is that it has to create a capability to predict the results given any set of starting conditions. Unfortunately, the works of Nester, van Langalaan and Lundgren, while being important additions to the literature, have not improved our predictability of foot and leg movements. I believe that the writing of Jones (Jones PC. Unwringing the Helix. Podiatry Management. 2011 Sep 1;30(7).) on trying to better elucidate the twisted plate concept is a good start in trying to model the restraints of motion but we need much more. Okita, et al. (Okita N, Meyers SA, Challis JH, Sharkey NA. Midtarsal joint locking: new perspectives on an old paradigm. Journal of Orthopaedic Research. 2014 Jan 1;32(1):110-5) has produced some additional interesting information about motions within the midtarsal joint during gait. It is hoped that the young readers of this will quit throwing rocks at those who have done their best to understand the midtarsal joint, from its restraining mechanism to its degrees of freedom of movement, and take up the challenge to improve our model that will lead to even better predictability and better treatment algorithms. As Dr. Wille challenged Dr. Root in 1975, the challenge remains to the entire profession. It's time for our biomechanics to contain mechanics and not just ‘bio.’

I find it interesting that Dr. Phillips spends so much time discussing the beliefs of Dr. Merton Root and his interactions with Dr. Root in his comment. My article was about midtarsal joint biomechanics, not about Dr. Root.

Contrary to the beliefs of Dr. Phillips, the research by Chris Nester, PhD and colleagues on midtarsal joint biomechanics is right on the money. It was done on live subjects in a weightbeaing setting. It wasn't done by artificially moving the forefoot on the rearfoot in the patterns that the examiner believes should exist in a non-weightbearing setting since this is what Dr. Root taught him. This is not how modern biomechanics research is performed.

Nester et al described the midtarsal joint using three reference axes to describe the spatial location of the midtarsal joint axis relative to the sagittal, frontal and transverse planes, which is consistent with the modelling of other joint axes within the human body within the international biomechanics community.   Modern biomechanics research is not performed by having an examiner moving the forefoot on the rearfoot to create an artificial "two-axis model" that corresponds to the primitive research of authors from over a half century ago. This "two-axis model, which  Dr. Phillips still defends, is nothing more than another podiatric myth that is still being perpetuated.

It is hoped that young readers will continue throwing rocks at those who have come before them in an attempt to break down the beliefs that are not based on scientific research and sound biomechanical principles. Hopefully, the young podiatrists of today will not form their opinions solely based on the beliefs of their podiatric teachers whom they admire and praise at the expense of their own objectivity regarding the scientific study of foot and lower extremity function.


I am sorry that Dr. Kirby failed to see the point of my reply to his criticisms of the traditional midtarsal joint axis model. My reply was to show that when there is good scientific technique which uses great technology, wrong conclusions can still be reached. Dr. Nester used great scientific technique but he came to the wrong conclusion: that we have to discard one theory in favor of another. This is not infrequent in the scientific literature through the ages. One could argue that Einstein's theories of relativity negated Newton's theory of gravity yet we continue to use Newton's theories to do a great number of common every day computations with a tremendous amount of accuracy. What I tried to point out was that we can accept the research by Dr. Nester on the axis of the midtarsal joint and show mathematically that it does not contradict the previous models. Such acceptance of both requires only simple linear algebra techniques. I also tried to point out that those who have taught the Manter model of the midtarsal joint axes have recognized that the model is not entirely correct. However, the model continues to suffice for a great many common and successful clinical practices. What I would like Dr. Kirby to do is show us how he teaches the kinesiology and the kinematics of the hip joint to young students in biomechanics, especially to students who may have only rudimentary knowledge and interest in three dimensional geometry, trigonometry, calculus and linear algebra. Does he teach that the hip joint has three axes of motion? Does he teach that the hip has a single instantaneous axis of motion or does he teach that the hip joint has an infinite number of axes? What are the advantages of each model of teaching? As to the ‘tossing rocks’ comment by Dr. Kirby, I point out that it is easy to criticize the weaknesses of any model and also to misrepresent the statements and attitudes of previous authors. It is much more difficult to build a new model. It is also commonly believed by many that newer literature is always better than older literature in the methodologies and therefore the conclusions. Nothing can be further from the truth. It is as important that the new literature be scrutinized for errors in methodology and analysis as the older literature. It is also important not to allow researchers to make leaps of faith in reaching their conclusions. The work by Nester on the midtarsal joint does just that. It uses good scientific technique and equipment, makes accurate computations, but then makes a leap of faith in stating that a previous model has to be discarded by misrepresenting it and not looking at how the new work actually can be integrated into previous models. When one starts to analyze the Nester conjecture and determine how it applies to daily activities, it soon becomes evident that a single axis of midtarsal joint motion would markedly limit the variety of terrains that one could walk over and the variety of activities one could engage in. On the other hand, a multiaxis midtarsal joint model markedly increases the available motions for the foot to undergo and increases the activities that one may engage in. I trust that Dr. Kirby will engage in the future in more academic discussions of this important subject.

I believe Dr. Phillips failed to see the point of Dr. Kirby's essay. The important point of Dr. Kirby's essay is that we should discard the notion that decrease in range of motion of the midtarsal joint seen with supination of the subtalar joint is not caused by changes in orientation of the axes of the midtarsal joint. This is the podiatric myth of the locking mechanism of the midtarsal joint. It is very important to understand how surgical procedures like the Evans calcaneal osteotomy can reduce motion of the midtarsal joint. The two axis theory cannot explain why there is a decrease in range of motion. The reason that the two axis theory cannot explain why there is a decrease in range of motion lies in the definition of axis of motion. An axis of motion is an imaginary line that describes the motion that occurs. An imaginary line cannot limit motion. Only ligaments and joint surfaces can limit motion. If we want to understand how midatarsal joint range of motion can be changed by position of the subtalar joint, or by the Evans calcaneal osteotomy, we need to look at how the joint surfaces limit motion and how changing the relationship of the calcaneal cuboid and talonavicular joint surfaces can alter motion and ligament tension. Eric Fuller, DPM
Back to Top