Strike type variation among Tarahumara Indians in m inimal sandals versus conventional running shoes

Daniel E.Lieberman

Department of Human Evolutionary Biology,Harvard University,Cambridge,MA 02138,USA

Strike type variation among Tarahumara Indians in m inimal sandals versus conventional running shoes

Daniel E.Lieberman

Department of Human Evolutionary Biology,Harvard University,Cambridge,MA 02138,USA

Purpose:This study exam ined variation in foot strike types,lower extremity kinematics,and arch height and stiffness among Tarahumara Indians from the Sierra Tarahumara,Mexico.

Methods:High speed video was used to study the kinematics of 23 individuals,13 who habitually wear traditional m inimal running sandals (huaraches),and 10 who habitually wearmodern,conventional running shoes w ith elevated,cushioned heels and arch support.Measurements of foot shape and arch stiffness were taken on these individuals plus an additional sample of 12 individuals.

Results:M inimally shod Tarahumara exhibitmuch variation w ith 40%primarily using m idfootstrikes,30%primarily using forefootstrikes,and 30%primarily using rearfootstrikes.In contrast,75%of the conventionally shod Tarahumara primarily used rearfootstrikes,and 25%primarily used m idfoot strikes.Individuals who used forefoot or m idfoot strikes landed w ith significantly more plantarflexed ankles,flexed knees,and flexed hips than runners who used rearfootstrikes.Footmeasurements indicate thatconventionally shod Tarahumara also have significantly less stiff arches than those wearing minimal shoes.

Conclusion:These data reinforce earlierstudies that there is variation among footstrike patterns among m inimally shod runners,butalso support the hypothesis that footstiffness and importantaspects of running form,including footstrike,differ between runners who grow up using m inimal versus modern,conventional footwear.

CopyrightⒸ2014,Shanghai University of Sport.Production and hosting by Elsevier B.V.A ll rights reserved.

Barefoot;Foot strike;M inimal shoes;Running;Tarahumara

1.Introduction

Recent studies of barefoot running have sparked interest in several aspects of running form,especially foot strike.Previous studies have shown that75%—90%of shod runners tend to rearfootstrike(RFS),landing fi rston the heel.1—3In contrast, several studies have reported that habitually barefoot runners are more likely to land w ith either a forefoot strike(FFS),in which the lateralmetatarsal heads fi rstmake contactw ith the ground,or with a midfootstrike(MFS),in which the heeland ball of the foot simultaneously contact the ground.4—6Other studies have found that habitually shod runners asked to run barefootoften sw itch from an RFS to an FFS when running on a hard surface such as asphalt.7One likely cause of these kinematic variations is the relationship between different foot strike types and vertical ground reaction forces(GRFv).RFSs generate an impactpeak,a rapid spike of the GRFvthatoccurs w ithin the fi rst 20—50 ms of stance,caused by the nearly instantaneous exchange ofmomentum between the ground and the effective mass(Meff)of the lower extremity that decelerates completely at the moment of impact.8In contrast, FFSs and some MFSs do not cause any measurable impact peak from more lower extrem ity compliance and lessMeff.6Accordingly,it is commonly hypothesized that barefootrunners are less likely to RFS when running long distances on hard,rough surfaces because repeated high,rapid impact peaks can be painful w ithout cushioning from a shoe’s heel.

Evidence that barefoot runners do not RFS as much as conventionally shod runners has received much attention because it suggests that running long distances on hard surfaces w ith an RFS may be uncommon from an evolutionary perspective even though cushioned shoe heels can make these impacts com fortable.Although some studies have questioned a relationship between repetitive stress injuries and repeated, rapid,and high impact rates,9,10others have found that the rate and magnitude of loading of impact peaks is associated w ith a range of injuries such as patello-femoral pain syndrome, medial tibial stress syndrome,Achilles tendonitis and plantar fasciitis.11—16In addition,there is limited evidence that habitually shod runners who FFS are less likely to incur repetitive stress injuries than those who RFS.17,18

Much attention has been paid to studies comparing strike types among habitually shod and barefoot runners,but there are reasons to question their relevance to most runners today trying to decide what footwear to use and whatkind of running form to adopt.First,all runners(including those who are barefoot)exhibitvariation in strike types depending on factors such as incline,speed,the characteristics of the substrate(e.g., hardness,roughness,and slipperiness),calf and foot muscle strength,and fatigue.6Future studies should focus more on these and other sources of variation among different populations.Second,strike type is only one non-independent aspect of running form that affects how repetitive forces are generated.Several studies have proposed that habitually barefoot runners tend to run w ith a high step frequency,little to no overstride(how far the ankle lands in frontof the knee), and w ith a relatively vertical trunk.4—6,19These aspects of form,which are not independent,may be relevant to injury prevention.Finally,the majority of runners today grow up wearing shoes,and they rarely run or walk barefoot for long distances.Most who choose not to use standard cushioned, elevated-heel shoes wear m inimal footwear that lack cushioned heels and arch supports.M inimal shoes are often marketed oxymoronically as“barefootshoes”,and actualbarefoot running is sometimes conflated spuriously w ith running in m inimal shoes.Despite much variation among m inimal shoes in terms of heel elevation,midsole thickness,and sole stiffness,these and other shoe components presumably affect proprioception from the plantar surface of the foot as well as how the foot lands and then functions during stance.The effects of m inimal footwear on running form and injury are poorly understood,but studies of western runners who have transitioned to minimalshoes suggest that they are more likely to RFS than actual barefoot runners,20,21w ith the possible effect of increasing the likelihood of certain injuries.22,23

This study aims to add to our understanding of the effects of footwear on variation in running form by examining a population of runners who traditionally wear m inimal footwear:the Tarahumara Native Americans from the Sierra Madre Occidental of northwestern Mexico(also known as the Sierra Tarahumara).The Tarahumara(self-identified as the Rara´muri)are one of several Native American groups that are famous for their tradition of running long distances in very rough terrain.Oral history and ethnographic accounts report that the Tarahumara used to run down prey such as deer and antelopes through endurance running.24—26This style of hunting,known as persistence hunting,takes advantage of two unique human abilities:to cool by sweating,and to run long distances at speeds that make quadrupeds gallop.Since quadrupeds cool by panting but cannot simultaneously pant and gallop,persistence hunting through endurance running can drive animals into a state of hyperthermia over long distances in hot,arid conditions.27—30The Tarahumara do not train in a conventional sense by running on a regular basis,but instead engage in long distance running several times a year by participating in therarajipari,an ancient ball game in which teams run long distances,often 75 km or more,while kicking and then chasing a small wooden ball.Tarahumara women compete in a slightly different long distance race known as theariwete,which uses a hoop rather than a ball,and typically involves distances of 40 km or less.The antiquity of therarajipariandariweteare unknown,but therarajipariis recorded by the earliest accounts of the Tarahumara,and probably dates back for many thousands of years.Recently, the Tarahumara have also started to compete in ultramarathons.A lthough the Tarahumara rarely if ever practice persistence hunting today,some older individuals report having done so when they were young,and long distance running remains an important partof their culture throughrarajiparis,ariwetes,and ultramarathons.These races have become well known because of the best-sellerBorn to Run,31but it is worth emphasizing that the Tarahumara are justone of many Native American groups that excelled at long distance running.32

A second reason to study variation in running kinematics among the Tarahumara is that they traditionally run in m inimal footwear.The Tarahumara inhabit an extremely mountainous region that is characterized by steep canyons and very rocky terrain.It is challenging to run barefootunder such conditions, and it is hardly surprising that Tarahumara traditionally run in sandals,known ashuaraches,that consist only of a piece of rawhide affixed fi rm ly to the sole of the footby a leather thong that goes between the fi rst and second toes as well as behind the ankle(Fig.1).According to one early account:they“are often stark naked except for a pair ofguarraches,or rawhide sandals,these protecting the feet from the fl int-like broken rocks of this part of the country,and w ithout which even their tough hides would soon be disabled”.33The Tarahumara now mostly fabricatehuarachesoles from car tires,but contemporaryhuarachesdiffer little from sandals recovered from archaeological sites,34,35indicating that the basic design has persisted for many thousands of years.Sim ilar sandals thatare suitable for running are common forms of footwear among many Native American groups,as are moccasins,which were also used for running.32,36Running sandals have recently become more common among western runners.

Fig.1.Huarachesworn by one of the m inimally shod participants(during warmup phase of the experiment).

A final reason to study the Tarahumara is thatmany aspects of their lifestyle are rapidly changing.Although some Tarahumara still wearhuarachesand other traditional clothes,as wellas grow corn and beans as they used to in isolated farms, many are becom ing westernized to varying extents.One dimension of this change is the increasingly common use of imported running shoes that have thick,cushioned,elevated heels,stiffened m idsoles,built-in arch supports,and toe springs.Even so,some of the Tarahumara who now wear modern running shoes still participate in traditional running events such as therarajipariand the ultramarathons.As a result,the Tarahumara represent a group in transition in terms of their footwear and running habits.

This study therefore collected data on variation in foot strikes as well as other aspects of running kinematics among Tarahumara who wearhuarachesas well as those who wear modern,conventional shoes.There are anecdotal reports that the Tarahumara predom inantly FFS,31,37but to date there have been no studies of their running kinematics.Three hypotheses are tested.The fi rst is that Tarahumara who primarily wearand run inhuarachesare more likely to FFS or MFS than those who wear and run in western shoes.Second,it is hypothesized that foot strike among m inimally shod Tarahumara is uncorrelated w ith speed,age,and body mass,but covaries w ith other kinematic variables purportedly associated w ith barefoot running,notably a high cadence,m inimal overstride,and a relatively vertical trunk.Finally,it is hypothesized that Tarahumara who wearhuaracheshave higher and stiffer arches than those who wear modern,supportive shoes.The basis for this hypothesis is that several features of modern running shoes,especially stiff midsoles and arch supports,likely decrease how much work the intrinsic muscles of the footdo, and thus lead to weaker arches that are therefore less stiff and more likely to be low.38—42

2.M aterials and m ethods

2.1.Sample

A sample of 35 Tarahumara(Table 1)were studied in the Sierra Tarahumara from the region around the Barranca de Urique and in the highlands between the Barrancas de Urique and Batopilas.Subjects were recruited by word of mouth w ith the help of a residentwho is well known to many Tarahumara and w ith a local Tarahumara teacher who speaks Rara´muri (the native language of the Tarahumara).Unfortunately,itwas difficult to recruita large sample of individuals because most traditional Tarahumara live in isolated farms,far from roads and towns.In addition,the Tarahumara tend to be reserved and wary of outsiders,and many potential subjects,especially women,declined requests to be videoed while running.

Of the 35 individuals studied,a combination kinematic and anthropometric data were collected from 23 individuals in December 2012.This sample includes 13 males(32.6±12.9 years,mean± SD)who wear onlyhuaraches(hereafter referred to as minimally shod Tarahumara),and 10 individuals (7 males,3 females,26.0±11.9 years,mean±SD)who wear western shoes(hereafter referred to as conventionally shod Tarahumara).Most of the conventionally shod individuals came from the town of Urique.Anthropometric data were collected in December 2013 from an additional sample of 12 individuals(10 who were minimally shod and 2 who wereconventionally shod).It was not possible to collect kinematic data from these individuals.

Table 1 Anthropometric and foot stiff differences between m inimally and conventionally shod Tarahumara(mean±SD).

None of the individuals measured had current lower extrem ity injuries,butkinematic data from three individuals(all conventionally shod)were notanalyzed for different reasons: one male ran in fl ip-flops;a second(1 female)was visibly distressed by the experiment,and ran in an awkward and evidently unnatural style;a third(1 female)was recorded w ith incorrect camera position.Accordingly,Table 2 presents data from 12 m inimally shod and eightconventionally shod individuals.All individualsgave their informed consentin Spanish orRaramuri according to protocols approved by Harvard University.

2.2.Anthropometrics and background information

Basic background and anthropometric information was collected from all participants including age,sex,height,body mass,and leg length(measured from the greater trochanter to the base of the heel).Participantswere asked to describe how far they travelevery day,whatkinds of physicalactivities they do on a regular basis,in what races they participate,and in what kind of footwear.Individuals were binned into fourcategories in terms of footwear usage,each assigned a rank-ordered numerical value:1)almost always shod in conventional shoes(less than 10%outdooractivity spent in minimalshoes orbarefoot); 2)usually shod in conventional shoes(wear shoes mostof the time,butdo athletic activities either in m inimal shoes or barefoot);3)m ixed(sometimes walk,run or do physicalactivity in conventional shoes and sometimes in m inimal shoes or barefoot);4)mostly minimally orshod barefoot(more than 80%of walking,running and physicalactivity done either in minimalshoes orbarefoot).They were also asked to describe any lower extremity injuries.Allquestions were asked in either Rara´muri or Spanish,and then translated into English.

Table 2 Kinematic differences between m inimally and conventionally shod Tarahumara(mean±SD).

Arch height and stiffness were assessed using an arch height index measurement system43that measures total foot length,the length of the foot from the back of the heel to the fi rst metatarsophalangeal joint(TFL),and the height of the dorsum of the foot at 50%of foot length(DH).Participants were measured both sitting and standing w ith 13 mm thick boards placed under the heeland the phalanges and metatarsal heads to enable the arch to move.Navicular height was measured by having individuals stand on a hard flatsurface(a concrete floor ora wooden board),placing a small ink mark on the navicular tuberosity,and then measuring the vertical distance between the ground and the mark using a rigid steel ruler accurate to 1 mm.Follow ing Zifchock and colleagues,43the arch height index(AHI)was calculated as DH/TFL both standing and seated;the arch stiffness index(ASI)was calculated as(body mass×0.4)/(AHIseated-AHIstanding).

2.3.Experimental trials

Participants were asked to wear whatever footwear they normally use,and to wear shorts or skirts that could be rolled up to reveal the knee.Reflective tape markers were placed on the follow ing locations on one side of the body:greater trochanter,the center of the knee(in between the lateral femoral epicondyle and the lateral tibial plateau),the lateral malleolus,the lateral surface of the 5th metatarsal head,and the lateralaspectof the tuber calcaneus.Participants were then photographed w ith a scale in lateraland frontalposition w ith a numeric identification.A ll participants were then instructed to run around an open field for approximately 5 m in at a pace they would choose when running a long distance.After the participant settled into a com fortable gait,step frequency was measured using an adjustable metronome(Matrix,New Market,VA,USA)fi tted w ith an earpiece.Preferred step frequency was considered to be the frequency attained once the cadence stayed constant for at least 1 min.Repeated measurements from the same subjects indicate thatstep frequency measurements are accurate to approximately 4 steps/m in.

Once each subject had warmed up,his or her running kinematics were then immediately recorded in lateral view on a trackway,approximately 15 m in length setup an a flat,grassfree and rock-free area,typical of the surfaces on which the Tarahumara normally walk and run in terms of surface hardness.A high-speed video camera(Casio EX-ZR100;Casio USA,Dover,NJ,USA)was positioned at 0.7 m height approximately 4 m lateral to the 10 m point on the track, providing an additional3—5 m of track beyond the field of the camera for subjects to run before decelerating.For each trial, individuals were asked to run down the trackway while looking forward and w ithout decelerating until they had passed a marker positioned approximately 3 m beyond the camera’s field of view.Participants were asked to run at different speeds they m ight use when running an ultramarathon orrarajipariwhile using a lightweight metronome eitherheld in the subject’s hand or clipped onto his/her clothes to control for step frequency.Although a range of step frequencies were measured(as part of a separate analysis,not reported here)only trials in which the subjects measured step frequencies were w ithin 5%of their preferred step frequency were used for analysis.In order to avoid having subjects alter their gait by stuttering their steps or overstriding as they passed the camera’s field of view,there was no landing target or region.If the marked foot did not land in front of the camera,the subject was asked to repeat the trial(w ithout explaining why),until a minimum of three trials were recorded.All sequences were recorded at 240 frames/s.

In order to avoid influencing how participants ran,no questions were asked about running form before or after the trials,and none were informed of the objectives of the experimentother than being told thatwe were measuring their feet and videoing them as they ran.

2.4.Kinematic analysis

A ll video sequences were converted to stacks of TIFF fi les and analyzed using ImageJ,version 1.46r(http://imagej.nih. gov/ij).Scale was determined for each subject using the distance between the markers on the lateral malleolus and knee. Running speed for each trial was quantified by measuring the horizontal translation of the greater trochanter between two homologous pointsduring a stride cycle(e.g.,toe-off to toe-off or foot strike to foot strike)relative to time(calculated from the number of frames divided by frame rate).Step frequency was calculated between the measured foot strike and the previous footstrike multiplied by the time sampled per frame.

Foot strike was measured using only sequences in which the markered foot landed in front of the camera perm itting a clear,non-parallaxed view of the lateral margin of the foot. Strike type was quantified using the angle of incidence(AOI), by measuring the orientation of the calcaneus and 5th metatarsal head markers relative to horizontal at the fi rst frame of contact m inus the same angle measured at foot flat.6,44AOIs greater than 1.0°were classified as FFS,AOIs between-1.0 and 1.0°were classified as MFS,and AOIs less than-1.0°were classified as RFS.Because AOI measurements can be affected by which frame is selected to represent the momentof contact,the reliability of assessing strike type by AOI was tested by classifying strike type independently through visual inspection of the video sequence.A strike was classified as an RFS if the vertical translation of the calcaneus marker stopped at the fi rst frame of ground contact while the 5th metatarsal head marker continued to move antero-inferiorly;as an FFS if the vertical translation of the 5th metatarsal head marker stopped at the fi rst frame of ground contact while the calcaneus marker continued to move inferiorly;and as an MFS if vertical translation of the calcaneus and 5th metatarsal head markers ceased simultaneously.The correlation between foot strike type determ ined by the two methods was 0.94 (p＜ 0.0001).Since the mean AOI only partially captures variation in strike types,a rank-order strike index was computed by assigning FFS,MFS,and RFS strikes scoresof1, 2,and 3,respectively,and then averaging.

As noted above,several other kinematic variables were measured at foot strike and at m idstance.Trunk angle was measured as the angle between the greater trochanter and the center of the neck relative to horizontal.Step frequency for each trial was quantified by measuring the time between two foot strike events(calculated from the number of frames divided by frame rate).Overstride(here defined as how far the ankle landed anterior to the knee)was measured as the angle of the lower leg(from the knee to the lateral malleolus) relative to earth horizontalat footstrike(an angle of 0°or less indicates no overstride,and higher angles indicate more overstride).An angular measurement of overstride was used instead of a linear measurementbecause there is less error in measuring the angle of the lower leg than in measuring the projected distance between the knee and ankle.Finally,to assess general lower extrem ity kinematics,hip angle was measured as the segment from the knee to the greater trochanter relative to horizontal;knee angle was measured as the angle between the lines from the knee to the greater trochanter and the knee to the lateral malleolus;and ankle angle was measured as the angle between the lines from the knee and to the lateral malleolus and from the lateral malleolus to the lateralmetatarsalhead.A llangles were measured by visual inspection using ImageJ.Since the data were collected under field conditions it is not possible to quantify accuracy,but reliability was assessed in two ways.First repeatability was quantified by taking the same set of measurements from one individualon five separate occasions.The average standard deviation was 0.32°w ith a range of 0.18°—0.49°.In addition,a test—retest sensitivity analysis conducted by taking all measurements tw ice from the same trial,yielding a correlation coefficient of 0.927.

2.5.Statistical analyses

Kinematic measurements were averaged foreach individual and compared among individuals and between minimally and conventionally shod groups primarily usingttests w ith footwear history(conventionalor m inimally shod)as the nom inal, dependentvariable.In addition,kinematic measurements were compared between individuals classified by strike type using ANOVA w ith modal strike type as the nominal,dependent variable.Mean individual values for strike type and other kinematic variables were also regressed against speed and step frequency.Strike type index and strike mode were compared between groups using a non-parametric Wilcoxon test.Effects were considered significant forp＜0.05.All analyses were done using JMP 5.0(SAS Institute,Cary,NC,USA).

3.Results

The two groups of Tarahumara,summarized in Table 1,did notdiffersignificantly in age,height,leg length,orbody mass, although as m ight be expected,the mean age of the conventionally shod Tarahumara subjects was nearly 8 years belowthe m inimally shod subjects(p=0.21,ttest).Footwear history,however,was very significantly different(p＜ 0.001, W ilcoxon test).This reflected the selection criteria used to define the two groups,w ith m inimally shod Tarahumara wearinghuarachesalmostexclusively,and the less traditional, conventionally shod individuals wearing them occasionally or rarely.Very few of the participants reported running barefoot as adults,although some of the m inimally shod Tarahumara said they would sometimes take off one or bothhuarachesfor kicking the ball during therarajipari,and children often run barefoot.

Although there is much variation,there were significant differences between the groups in terms of strike types,as summarized in Table 2.Among the m inimally shod Tarahumara,40%had a modal MFS strike type,30%had a modal FFS strike type,and 30%had a modalRFS strike type.Among the conventionally shod Tarahumara,75%had an RFS modal strike type,and 25%had an MFS modal strike type.As Fig.2A illustrates,this difference was reflected in mean strike type,which averaged 2.04 for the m inimally shod Tarahumara and 2.69 for the conventionally shod Tarahumara reflecting the predom inance of MFS landings among the former and RFS landings among the latter(p=0.045,W ilcoxon test).AOIs (Table 2)also indicate that the ankle was significantly more dorsiflexed in the conventionally shod versus m inimally shod groups(p=0.04,ttest).Speeds used ranged between 2.3 m/s and 4.8 m/s,but as Fig.2B shows,there was no significant correlation between speed and AOI for subject averages (r=0.04;p=0.83)or for all trials(r=0.02;p=0.85),nor did it correlate significantly w ith other anthropometric variables.Strike type,however,did correlate significantly w ith step frequency(r=0.47;p=0.03,ANOVA),w ith individuals who used higher step frequencies being more likely to FFS or MFS.

Given the high degree of variation w ithin the m inimally shod group,which included individuals who used RFS,MFS, and FFS landings,there were not many significant kinematic differences between the groups.A lthough the conventionally shod Tarahumara had a tendency to have lower preferred step frequencies,neither preferred step frequency nor the step frequency used during the trials differed significantly.Speed also did not differ between the groups.Although ankle,knee, hip,and trunk angles at foot strike and m idstance were not significantly different between the conventionally and m inimally shod Tarahumara,the degree of overstride as measured by the angle of the tibia at foot strike was significantly lower in minimally shod Tarahumara(p=0.003,ttest).In other words,the minimally shod Tarahumara were more likely to land w ith their ankles below the knee and w ith a more vertical lower leg.Fig.3 shows that when kinematic differences between individuals are plotted by mean strike type(as determined from the AOI),the rearfoot strikers had significantly more dorsiflexed ankles(p=0.012,ANOVA),more extended knees(p=0.007,ANOVA),and less flexed hips(p=0.006, ANOVA)at foot strike,but did not differ in terms of trunk angle at foot strike(p=0.610,ANOVA),nor in any of these kinematic variables atmidstance(results notshown).FFS and MFS did not differ significantly in any kinematic variables measured.

Finally,variations in arch heightand stiffness,summarized in Table 1,indicate that there were no significant differences between groups for absolute arch height(as measured by navicular height),relative arch height(navicular height standardized by truncated foot length)or the arch height index (measured during standing).However,arch stiffness was almost tw ice as great(p=0.003,ttest)among the m inimally shod Tarahumara who wearhuarachesthan among conventionally shod individuals.Two of the participants,both conventionally shod,had bilateralpes planus.

4.Discussion

Fig.2.Foot strike differences between m inimally and conventionally shod Tarahumara.(A)strike type index as determ ined by angle of incidence.Box and whiskers indicate mean,±1 and±2SD.(B)relationship between speed and angle of incidence.Closed circles,minimally shod subjects;open squares, conventionally shod subjects.

Fig.3.Kinematic differences at foot strike between m inimally and conventionally shod Tarahumara divided by footstrike type.Box and whiskers indicate mean, ±1 and±2SD.*signifi cant between FFS and RFS.Abbreviations:FFS=forefoot strike;MFS=forefoot strike;RFS=rearfoot strike;ns=not significant.

Experiments such as those reported here necessarily sample only a portion of the variation that actually exists.With that caveat in mind,minimally shod Tarahumara runners appear to be best characterized as midfoot strikers who also employ forefoot and rearfoot strikes.These data therefore partially support but also modify earlier anecdotal reports that Tarahumara runners who usehuarachesprimarily FFS.The variation w ithin the minimally shod sample,however,is substantial,making any simple characterization of their strike type problematic.In contrast,Tarahumara who have adopted western running shoes w ith cushioned,elevated heels,stiff soles,partial arch supports,toe springs and other features common in most athletic shoes are almost entirely rearfoot strikers,much like habitually shod runners who have been measured in countries such as the USA.2,3The contrast between minimally shod and conventionally shod Tarahumara therefore supports the general observation that experienced runners who are habitually barefootorm inimally shod tend to land w ith a flatter foot and RFS less often than runners who habitually use modern running shoes that cushion and elevate the heel.4—6,45,46

There is much more to running form than strike type. Although there were no statistically significant differences in step frequency as well as the angles of the ankle,knee,hip, and trunk between the m inimally and conventionally shod individuals,m inimally shod Tarahumara had significantly less overstride(as measured by the orientation of the tibia at landing),which means that they typically land w ith theirankle nearly below the knee.In addition,when subjects were grouped by strike type,individuals who used FFS and MFS landings(all but one of which was m inimally shod)had significantly more plantarflexed ankles,flexed knees,and flexed hips at the moment of foot strike than individuals who use RFS landings.These kinematic differences were no longer evidentby m idstance.These data corroborate previous reports that barefoot and m inimally shod runners tend to differ from habitually shod runners in a number of aspects of running form.19The data also highlight the general kinematic similarity between forefoot and midfoot striking,both of which differ from rearfoot striking in a number of respects,perhaps the most important being less overstride in which the ankle lands nearly below the knee at the momentof strike,providing more limb compliance at the ankle and knee.

Previous studies have found that shod runners are more likely to FFS at higher speeds,47and one study found an increased use of FFS athigher speeds among the Daasenech,a habitually barefoot population from northern Kenya that does not run very much,and which lives in a very sandy habitat.48In contrast,this study found no effectof speed on strike type. One explanation for this result could be that the range of speeds employed was not great(2.3—4.8 m/s),w ith most runners choosing approximately 3.3—3.9 m/s.In addition, many of the runners already used FFS and MFS landings.It was notpossible to test for the effectof sex on strike type,but there was no effect of age,body mass,preferred step frequency on strike type variation.Future research is therefore needed to understand when and why runners who are m inimally shod or barefoot adopt different strike types.In this regard,additional variables to consider are how strong the runner is,especially in terms of the triceps surae and the foot muscles,the effects of distance and fatigue,and the influence of variations in the hardness,roughness,or slipperiness of thesubstrate.Since m inimal shoes such ashuarachespresumably allow less proprioception than being barefootbutconsiderably more than standard running shoes(a hypothesis that merits careful testing),it is reasonable to hypothesize that substrate characteristics have less of an effect on strike type choice among minimally shod than barefoot runners,but that other factors related to the skill of running long distances such as overstride,cadence,and posture remain just as important, perhaps even more so when running very long distances.

Although the focus of this study was on strike type,the results presented here also provide evidence for an effect of footwear on arch morphology and function.Most importantly,the m inimally shod Tarahumara had significantly stiffer arches as measured by differences in the arch stiffness index.The index measure only static stiffness during sitting and standing rather than during walking or running,but nonetheless suggests that the Tarahumara who wearhuaracheshad stronger intrinsic muscles that lead to a stiffer longitudinal arch.Further data are necessary to measure actual foot strength,but this result accords w ith other studies which have found thathabitually barefootor m inimally shod individuals have less variation in arch shape w ith a lower likelihood ofpes planus,41and that individuals who wear m inimal shoes develop more longitudinal arch strength as well as stiffer arches.40

Many lim itations caution against over interpreting the results of this study.Most obviously,sample sizes were necessarily small given the challenges of recruiting Tarahumara to participate.Many individuals simply did not want to be measured and videoed.Another non-trivial deficiency is that this study looked only at individuals running on a trackway under conditions that controlled for justa few variables(e.g., incline and step frequency)that were comparatively unchallenging relative to the demanding and extremely varied conditions under which Tarahumara usually run.The canyons of the Sierra Tarahumara are among the most rugged landscapes in the world because they are characterized by extreme,steep changes in elevation,lots of rocks,and very few flat areas. The trackways on which participants were recorded facilitated comparisons of runners,but they were easy and relatively com fortable running surfaces for the Tarahumara.In addition, the Tarahumara rarely go for short runs and they do not train, but instead are celebrated for their abilities and proclivities to run very long distances.Thus in order to characterize how they really run(i.e.,their full range of variation),it would be necessary to record individuals after many kilometers of running on conditions vastly different and more challenging than those used here.A lthough it is possible that runners inhuarachesare more likely to FFS when running long distances on rocky terrain to m inimize impact loading,it is also possible that the increased eccentric contractions of the plantarflexor muscles required during forefoot striking is too tiring over very long distances,favoring the adoption of m idfoot strikes.In all likelihood,Tarahumara runners probably use all kinds of strike types over the course of a 50—100km race.Future studies are currently planned to quantify this variation.

Finally,it is worth considering the relevance of these results for the majority of runners who grow up wearing shoes,rarely if ever run ultramarathons,and are habituated to conventional running shoes w ith cushioned,elevated heels,stiff m idsoles, orthotics,and toe-springs.Evidence that traditional Tarahumara who wearhuarachesmostly avoid RFS landings on flat surfaces atmoderate speeds is hardly justification for someone to sw itch to minimal shoes and stop heelstriking.It is possible that people who grow up wearing conventional shoes have weaker feet,and unless they are used to forefoot or m idfoot striking,they likely have weaker calf muscles less able to handle the additionaleccentric loading these styles of running demand.That said,there is some utility to studying how the Tarahumara run,and how using minimal versus conventional shoes affects their running as wellas their feetbecause a large percentage of conventionally shod people develop repetitive stress injuries and foot problems.49As the results presented here show,Tarahumara who run in conventional shoes tend to RFS like people all over the world who wear sim ilar shoes, and they apparently have more compliant arches than those who wearhuaraches.This study did not collect data on injuries,but several studies have shown how and why runners who generate higher and faster rates of impact loading are more likely to develop a suite of repetitive stress injuries.9,10,17In this regard,evidence that Tarahumara who wearhuarachesare less likely to RFS makes sense because itenables them to avoid painful and potentially damaging impacts rather than use cushioned heels to merely slow the rate of impact loading. To be sure,habitually shod runners who w ish to adoptm inimal shoes or change their kinematics,should do so cautiously, gradually,and properly(e.g.,w ithout overstriding),allow ing the body to adaptappropriately.However,there may be some w isdom in using traditional kinds of footwear and learning traditional ways of running.

Acknow ledgm ents

For help conducting research in the Copper Canyons I am especially grateful to M ickey Mahaffey.Additional field assistance was provided by Flora Ayala Frias,Aaron Baggish, Sara Del Castillo,Ignacio Iglessis,Stephanie Mahaffey,Jennifer Neary,and Evan Sofro.Funding was provided by a grant from the American School of Prehistoric Research(Harvard University).For constructive comments and discussions I thank Eric Castillo,Heather Dingwall,Herman Pontzer,and two anonymous referees.

1.Kerr BA,Beauchamp L,Fisher V,Neil R.Footstrike patterns in distance running.In:Nigg BM,Kerr B,editors.Biomechanical aspects of sport shoes and playing surfaces.Calgary:University of Calgary Press;1983. p.135—42.

2.Hasegawa H,Yamauchi T,Kraemer WJ.Foot strike patterns of runners at the 15-km point during an elite-level half marathon.J Strength Cond Res2007;21:888—93.

3.Larson P,Higgins E,Kam inski J,Decker T,Preble J,Lyons D,etal.Foot strike patterns of recreationaland sub-elite runners in a long-distance road race.J Sports Sci2011;29:1665—73.

4.De Wit B,De Clercq D,Aerts P.Biomechanical analysis of the stance phase during barefoot and shod running.J Biomech2000;33:269—78.

5.Squadrone R,GallozziC.Biomechanicaland physiologicalcomparison of barefoot and two shod conditions in experienced barefoot runners.J Sports Med Phys Fitness2009;49:6—13.

6.Lieberman DE,Venkadesan M,Werbel WA,Daoud AI,D’Aeandrea S, Davis IS,et al.Foot strike patterns and collision forces in habitually barefoot versus shod runners.Nature2010;463:531—5.

7.Nigg BM.Biomechanics ofsports shoes.Calgary:Topline Printing;2010. p.300.

8.Whittle MW.Generation and attenuation of transient impulsive forces beneath the foot:a review.Gait Posture1999;10:264—75.

9.Nigg BM,Wakeling JM.Impact forces and muscle tuning:a new paradigm.Exerc Sport Sci Rev2001;29:37—41.

10.Nigg BM,Cole GK,Bru¨ggemann G-P.Impact forces during heel-toe running.J Appl Biomech1995;11:407—32.

11.Hreljac A.Impact and overuse injuries in runners.Med Sci Sports Exerc2004;36:845—9.

12.Hreljac A,MarshallRN,Hume PA.Evaluation of lowerextremity overuse injury potential in runners.Med Sci Sports Exerc2000;32:1635—41.

13.M ilner CE,Ferber R,Pollard CD,Hamill J,Davis IS.Biomechanical factors associated w ith tibial stress fracture in female runners.Med Sci Sports Exerc2006;38:323—8.

14.Pohl MB,Ham ill J,Davis IS.Biomechanical and anatom ic factors associated with a history of plantar fasciitis in female runners.Clin J Sport Med2009;19:372—6.

15.Davis IS,Bowser B,Mullineaux D.Do impacts cause running injuries?A prospective investigation.ASB;2010.Available at:http://www.asbweb. org/conferences/2010/abstracts/472.pdf[accessed 03.04.2014].

16.Zadpoor AA,Nikooyan AA.The relationship between lower-extremity stress fractures and the ground reaction force:a systematic review.Clin Biomech2011;26:23—8.

17.Daoud IA,Geissler GJ,Wang F,Saretsky J,Daoud YA,Lieberman DE. Foot strike and injury rates in endurance runners:a retrospective study.Med Sci Sports Exerc2012;44:1325—34.

18.Goss DL,Gross MT.Relationships among self-reported shoe type,footstrike pattern,and injury incidence.US Army Med Dep J2012:25—30.

19.Lieberman DE.What we can learn about running from barefoot running:an evolutionary medical perspective.Exerc Sport Sci Rev2012;40:63—72.

20.Bonacci J,Saunders PU,Hicks A,Rantalainen T,Vicenzino BG, Spratford W.Running in a m inimalistand lightweightshoe is not the same as running barefoot:a biomechanical study.Br JSports Med2013;47:387—92.

21.W illy RW,Davis IS.Kinematic and kinetic comparison of running in standard and m inimalist shoes.Med Sci Sports Exerc2013;46:318—23.

22.Salzler M J,Bluman EM,Noonan S,Chiodo CP,de Asla RJ.Injuries observed in m inimalist runners.Foot Ankle Int2012;33:262—6.

23.Ridge ST,Johnson AW,M itchell UH,Hunter I,Robinson E,Rich BS, et al.Foot bone marrow edema after a 10-wk transition to minimalist running shoes.Med Sci Sports Exerc2013;45:1363—8.

24.Bennett W,Zingg R.The Tarahumara:an Indian tribe of Northern Mexico.Chicago:University of Chicago Press;1935.p.412.

25.Pennington C.The Tarahumara of Mexico.Salt Lake City:University of New Mexico Press;1969.p.267.

26.Kennedy JG.Tarahumara of the Sierra Madre.Pacifi c Grove,CA:Asilomar Press;1978.p.299.

27.Carrier DR.The energetic paradox of human running and hom inid evolution.Curr Anthropol1984;24:483—95.

28.Bramble DM,Lieberman DE.Endurance running and the evolution of Homo.Nature2004;432:345—52.

29.Lieberman DE,Bramble DM.The evolution of marathon running capabilities in humans.Sports Med2007;37:288—90.

30.Liebenberg L.Persistence hunting by modern hunter-gatherers.Curr Anthropol2006;47:1017—26.

31.M cDougall C.Born to run:a hidden tribe,superathletes,and the greatest race the world has never seen.New York:Knopf;2009.p.304.

32.Nabokov P.Indian running:native American history and tradition.Santa Barbara:Capra Press;1981.p.207.

33.Schwatka F.In the land of the cliffand cave dwellers.New York:Cassell; 1893.p.313.

34.Cressman LS.The sandal and the cave:the Indians of Oregon.Corvallis, OR:Oregon Historical Society;1981.p.81.

35.Kuttruff JT,DeHart SG,O’Brien M J.7500 years of prehistoric footwear from Arnold Research Cave,M issouri.Science1998;281:72—5.

36.Kroeber AL.Handbook of the Indians of California.Washington DC: Bureau of American Ethnology;1925.p.995.

37.Jurek S.Eat and run:my unlikely journey to ultramarathon greatness. New York:Houghton-M iffl in;2012.p.260.

38.Perl DP,Daoud AI,Lieberman DE.Effects of footwear and strike type on running economy.Med Sci Sports Exerc2012;44:1335—43.

39.Robbins SE,Hanna AM.Running-related injury prevention through barefootadaptations.Med Sci Sports Exerc1987;19:148—56.

40.Bruggemann GP,PotthastW,Braunstein B,Niehoff A.Effectof increased mechanical stimuli on foot muscles functional capacity.In:Proceedings International Society of Biomechanics XXthCongress.Cleveland: American Society of Biomechanics;2005.p.553.

41.D’Aouˆt K,Pataky TC,De Clercq D,Aerts P.The effects of habitual footwear use;foot shape and function in native barefoot walkers.Footwear Sci2009;1:81—94.

42.Richards CE,Magin PJ,Callister R.Is your prescription of distance running shoes evidence-based?Br J Sports Med2009;43:159—62.

43.Zifchock RA,Davis I,Hillstrom H,Song J.The effectof gender,age,and lateral dom inance on arch height and arch stiffness.Foot Ankle Int2006;27:367—72.

44.A ltman AR,Davis IS.A kinematic method for footstrike pattern detection in barefoot and shod runners.Gait Posture2012;35:298—300.

45.Divert C,Mornieux G,Freychat P,Baly L,Mayer F,Belli A.Barefootshod running differences:shoe or mass effect.Int J Sports Med2008;29:512—8.

46.Jenkins DW,Cauthon DJ.Barefoot running claims and controversies:a review of the literature.J Am Podiatr Med Assoc2010;101:231—46.

47.Keller TS,Weisberger AM,Ray JL,Hasan SS,Shiavi RG,Spengler DM. Relationship between vertical ground reaction force and speed during walking,slow jogging,and running.Clin Biomech1996;11:253—9.

48.Hatala KG,Dingwall HL,Wunderlich RE,Richmond BG.Variation in footstrike patterns during running among habitually barefootpopulations.PLoSOne2013;8:e52548. http://dx.doi.org/10.1371/journal.pone. 0052548.

49.van Gent RN,Siem D,van M iddelkoop M,van Os AG,Bierma-Zeinstra SM,Koes BW.Incidence and determ inants of lower extrem ity running injuries in long distance runners:a systematic review.Br J Sports Med2007;41:469—80.

Received 2 September 2013;revised 9 February 2014;accepted 10 March 2014

E-mail address:danlieb@fas.harvard.edu

Peer review under responsibility of Shanghai University of Sport

2095-2546/$-see front matter CopyrightⒸ2014,Shanghai University of Sport.Production and hosting by Elsevier B.V.A ll rights reserved. http://dx.doi.org/10.1016/j.jshs.2014.03.009