JCLB-03755; No of Pages 5 Clinical Biomechanics xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech

Effects of barefoot and barefoot inspired footwear on knee and ankle loading during running Jonathan Sinclair Centre for Applied Sport and Exercise Sciences, University of Central Lancashire, Lancashire, UK Darwin Building 217, Centre for Applied Sport Exercise and Nutritional Sciences, School of Sport Tourism and Outdoors, University of Central Lancashire, Preston, Lancashire PR1 2HE, UK

a r t i c l e

i n f o

Article history: Received 16 December 2013 Accepted 18 February 2014 Keywords: Barefoot Minimalist footwear Achilles tendon load Patellofemoral contact

a b s t r a c t Background: Recreational runners frequently suffer from chronic pathologies. The knee and ankle have been highlighted as common injury sites. Barefoot and barefoot inspired footwear have been cited as treatment modalities for running injuries as opposed to more conventional running shoes. This investigation examined knee and ankle loading in barefoot and barefoot inspired footwear in relation to conventional running shoes. Method: Thirty recreational male runners underwent 3D running analysis at 4.0 m·s−1. Joint moments, patellofemoral contact force and pressure and Achilles tendon forces were compared between footwear. Findings: At the knee the results show that barefoot and barefoot inspired footwear were associated with significant reductions in patellofemoral kinetic parameters. The ankle kinetics indicate that barefoot and barefoot inspired footwear were associated with significant increases in Achilles tendon force compared to conventional shoes. Interpretation: Barefoot and barefoot inspired footwear may serve to reduce the incidence of knee injuries in runners although corresponding increases in Achilles tendon loading may induce an injury risk at this tendon. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction Recreational runners are known to be susceptible to injuries; 19.4–79.3% of all who participate in running activities will suffer from a chronic pathology over the course of one year (van Gent et al., 2007). The knee and ankle joint structures have been demonstrated to be the most common injury sites and are associated with one-fifth of all running-related injuries (van Gent et al., 2007). Recently barefoot (BF) running has been the focus of much attention in footwear biomechanical research. The rise in popularity of barefoot locomotion is borne out of the hypothesis that running without shoes is associated with a reduced incidence of chronic running injuries (Lieberman et al., 2010). The rationale behind this notion is that the non-rearfoot strike pattern typically associated with barefoot locomotion serves to attenuate the magnitude and rate of the impact experienced by the musculoskeletal system as a result of the foot striking the ground (Sinclair et al., 2013a). Taking into account the barefoot movement's recent rise in popularity and potential propensity to reduce injury aetiology, shoes have been designed in an attempt to transfer the prospective benefits of barefoot movement into a shod condition

E-mail address: [email protected]

(Sinclair et al., 2013a). Numerous barefoot inspired (BFIS) footwear models are currently available and vary in design characteristics from minimalistic e.g. Inov-8 Evoskin and Vibram Five Fingers to more structured designs which offer some midsole interface e.g. Nike Free (Sinclair et al., 2013a). A number of studies have investigated the joint kinetics/kinematics of running BF and BFIS in relation to conventional running shoes. Sinclair et al. (2013a) showed that running BF was associated with increases in vertical rates of loading and ankle plantarflexion angle during the stance phase of running. Sinclair et al. (2013b) examined several BFIS models in relation to both BF and conventional running shoes. It was demonstrated that BF and minimalist BFIS were associated with increases in vertical rates of loading, ankle plantarflexion and also peak rearfoot eversion. Bonacci et al. (2013a) examined joint kinetics during BF and BFIS running, and it was demonstrated that BF and BFIS locomotion are associated with reduced knee extensor and increased ankle plantarflexion moments. Similarly in examinations of rear and mid/forefoot strike patterns, mid/forefoot runners have also been shown to exhibit increases in Achilles tendon force (ATF) and reductions in patellofemoral contact force (PTF) and pressure (PP) in comparison to those exhibiting a rearfoot strike pattern (Kulmala et al., 2013). Similarly, Bonacci et al. (2013b) showed that running without shoes was linked to significant reductions in PTF and PP compared to running footwear. It can be speculated based on these prevalent findings in

http://dx.doi.org/10.1016/j.clinbiomech.2014.02.004 0268-0033/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article as: Sinclair, J., Effects of barefoot and barefoot inspired footwear on knee and ankle loading during running, Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2014.02.004

2

J. Sinclair / Clinical Biomechanics xxx (2014) xxx–xxx

relation to joint kinetics at the knee and ankle joints that BF and BFIS may alter loading patterns at these joints, however currently there is a paucity of information investigating PTF and PP at the knee and ATF patterns of the ankle during BFIS in comparison to conventional running footwear. The aim of the current investigation was therefore to determine whether running in BF and BFIS footwear caused different levels of PTF and PP at the knee and ATF at the ankle in comparison to conventional running trainers. Specifically it was hypothesized that running BF and in minimalist BFIS would be associated with reduced PTF and PP in relation to conventional footwear. 2. Methods 2.1. Participants Thirty recreational male runners, who engage in running activities a minimum of three times per week, were recruited for this investigation. The participants' mean characteristics were: age = 26.21, SD = 5.52 years, body mass = 73.45, SD = 6.00 kg, height = 1.77, and SD = 0.09 m. Participants were all free from pathology at the time of data collection and provided informed consent in written form. The procedure for the study was approved by a university ethical panel. 2.2. Procedure The participants completed ten running trials over a 22 m walkway at 4.0 m·s−1 in a biomechanics laboratory. A maximum deviation from the pre-determined velocity of 5% was allowed. The participants struck an embedded piezoelectric force platform (Kistler Instruments, Model 9281CA; Winterthur, Switzerland, Dimensions = 0.6 × 0.4 m) sampling at 1000 Hz with their right foot (Sinclair et al., 2013c). Running velocity was monitored using infra-red timing gates (SmartSpeed Ltd, Cardiff, UK). The stance phase of the running cycle was delineated as the time over which a minimum of 20 N vertical force was applied to the force platform (Sinclair et al., 2013d). An eight camera optoelectric motion capture system was used to collect 3D kinematics. The synchronised kinematic and ground reaction force data were obtained using Qualisys Track Manager software (Qualisys Medical AB, Goteburg, Sweden) with a capture frequency of 250 Hz. The calibrated anatomical system technique (CAST) was utilised to quantify joint kinematics (Cappozzo et al., 1995). To define the anatomical frames of the right foot, shank and thigh, retroreflective markers were positioned onto the calcaneus, 1st and 5th metatarsal heads, medial and lateral malleoli, medial and lateral epicondyle of the femur and greater trochanter. Rigid carbon-fibre tracking clusters comprising of four retroreflective markers were positioned onto the thigh and shank segments and secured using tape. The foot was tracked using the calcaneus, first and fifth metatarsal markers. Static calibration trials (not normalized to standing posture) were obtained with the participant in the anatomical position in order for the positions of the anatomical markers to be referenced in relation to the tracking clusters/markers. Separate static trials were obtained for each footwear condition. 2.3. Data processing GRF and 3D kinematic data were filtered at 50 and 12 Hz respectively using a low pass Butterworth 4th order zero-lag filter and subsequently analysed using Visual 3D (C-Motion, Germantown, MD, USA). 3D kinematics of the knee and ankle were calculated using an XYZ Cardan sequence of rotations (where X = sagittal plane; Y = coronal plane and Z = transverse plane). Kinematic curves were normalized to 100% of the stance phase then processed trials were averaged. Joint kinetics were computed using Newton–Euler

inverse-dynamics. To quantify net joint moment anthropometric data, ground reaction forces and angular kinematics were used. The net joint moments were normalized to body mass and kg·N·m. A previously utilized algorithmic model was used to determine patellofemoral contact force and pressure (Ward and Powers, 2004). This algorithm has been utilized previously to resolve differences in PTF and PP when wearing different footwear (Bonacci et al., 2013b; Ho et al., 2012; Kulmala et al., 2013) and between those with and without patellofemoral pain (Heino and Powers, 2002). Patellofemoral joint contact force (B·W) during running was then estimated as a function of knee flexion angle (fa) and knee extensor moment (ME) according to the biomechanical model described by Ho et al. (2012). Firstly, an effective moment arm of the quadriceps muscle (mq) was calculated as a function of knee flexion angle using non-linear equation, which is based on cadaveric information presented by van Eijden et al. (1986): 3

2

mq ¼ 0:00008f a –0:013f a þ 0:28fa þ 0:046:

ð1Þ

Quadriceps force (QF) was then calculated using the below formula: QF ¼ ME=mq:

ð2Þ

PTF was estimated using the QF and a constant (K): PTF ¼ QF K:

ð3Þ

The constant was described in relation to the fa using a curve fitting technique based on the non-linear equation described by van Eijden et al. (1986):   2 2 K ¼ 0:462 þ 0:00147f a –0:0000384f a =ð1–0:0162fa 2

3

þ0:000155 f a –0:000000698f a Þ:

ð4Þ

PP (MPa) was calculated as a function of the PTF divided by the patellofemoral contact area. The contact area was described in accordance with the Ho et al. (2012) recommendations by fitting a second-order polynomial curve to the data of Powers et al. (1998) who documented patellofemoral contact areas at varying levels of knee flexion (83 mm2 at 0°, 140 mm2 at 15°, 227 mm2 at 30°, 236 mm2 at 45°, 235 mm2 at 60°, and 211 mm2 at 75° of knee flexion). PP ¼ PTF=contact area

ð5Þ

Achilles tendon force (ATF) (B·W) was determined by dividing the plantarflexion moment (MPF) by the estimated Achilles tendon moment arm (mat). The moment arm was quantified as a function of the ankle sagittal plane angle (ak) using the procedure described by Self and Paine (2001): ATF ¼ MPF=mat

ð6Þ

2

mat ¼ −0:5910 þ 0:08297ak–0:0002606ak :

ð7Þ

PTF and ATF loading rates (B·W·s−1) were calculated as a function of the change in tendon force from initial contact to peak force divided by the time to peak force. 2.4. Footwear The experimental footwear utilized during this study consisted of a Saucony Pro Grid Guide II, Vibram Five Fingers, Inov-8 Evoskin and Nike Free 3.0 in sizes 8–10 men's UK (Fig. 1). The Saucony footwear were selected to provide a good representation of commercially available conventional footwear as they provide both midsole cushioning and medial support. The Nike Free 3.0, Vibram Five Fingers and Inov-8

Please cite this article as: Sinclair, J., Effects of barefoot and barefoot inspired footwear on knee and ankle loading during running, Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2014.02.004

J. Sinclair / Clinical Biomechanics xxx (2014) xxx–xxx

3

Fig. 1. Experimental footwear (a. =conventional, b. Nike Free 3.0, c. Vibram Five Fingers and d. Inov-8 Evoskin).

Evoskin were chosen for similar reasons to represent the variety of commercially available BFIS footwear. 2.5. Statistical analysis Differences in knee and ankle loading parameters across the footwear conditions were examined using one-way repeated measures ANOVA. The alpha criterion for statistical significance adjusted to P = 0.008 using a Bonferroni correction to control type I error. Effect sizes were calculated using an Eta2 (η2). Post-hoc pairwise comparisons were conducted on all significant main effects. The data was screened for normality using a Shapiro–Wilk which confirmed that the normality assumption was met. All statistical actions were conducted using SPSS 21.0 (SPSS Inc., Chicago, USA). 3. Results Fig. 2 presents the knee and ankle joint angles/kinetics obtained as a function of footwear and Table 1 exhibits the discrete kinematic parameters. The results indicate that footwear significantly influenced both knee and ankle kinetic parameters. 3.1. Knee load A significant main effect P b 0.006, η2 = 0.48 was found for ME (Table 1; Fig. 1b). Post-hoc analysis showed that ME was significantly greater in the conventional footwear in comparison to BF and Inov-8. A significant main effect P b 0.006, η2 = 0.52 was also observed for PTF (Table 1; Fig. 2c). Post-hoc analysis showed that PTF was significantly greater in the conventional and Nike Free footwear in comparison to BF, Vibram Five Fingers and Inov-8. A further main effect P b 0.006, η2 = 0.41 was found for PP (Table 1; Fig. 1d). Post-hoc analysis showed that PP was significantly greater in conventional and Nike Free footwear in comparison to BF and Vibram Five Fingers. A final main effect P b 0.006, η2 = 0.40 was also shown for PTF load rate (Table 1). Post-hoc analysis showed that PTF load rate was

Fig. 2. Knee and ankle kinetics and kinematics as a function of footwear, black = BF, Grey = Vibram Five Fingers, Dash = Nike Free, Dot = conventional, black outline = Inov-8 (a = knee angle, b = knee moment c = PTF, d = PP, e = ankle angle, f = ankle moment, g = ATF).

significantly greater in the conventional footwear in comparison to BF and Vibram Five Fingers. 3.2. Ankle load A significant main effect P b 0.006, η2 = 0.43 was found for peak MPF (Table 1; Fig. 2f). Post-hoc analysis showed that MPF was significantly greater in BF and Inov-8 compared to the conventional and Nike Free footwear. A significant main effect P b 0.006, η2 = 0.48 was also found for ATF (Table 1; Fig. 2g). Post-hoc analysis showed that ATF was significantly greater in BF and Inov-8 compared to the conventional and Nike Free footwear. A final main effect P b 0.006, η2 = 0.49 was found for ATF load rate (Table 1). Post-hoc analysis showed that ATF load rate was significantly greater in BF compared to the conventional and Nike Free footwear.

Please cite this article as: Sinclair, J., Effects of barefoot and barefoot inspired footwear on knee and ankle loading during running, Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2014.02.004

4

J. Sinclair / Clinical Biomechanics xxx (2014) xxx–xxx

Table 1 Mean & SD, joint kinetic information as a function of footwear. Barefoot

Peak ME (kg·N·m) PTF (B·W) Time to PTF (s) PP (Mpa) PTF load rate (B·W·s−1) Peak MPF (kg·N·m) ATF (B·W) Time to ATF (s) ATF load rate (B·W·s−1)

Vibram Five Fingers

Inov-8

Mean

SD

Mean

SD

Mean

SD

Mean

Nike Free SD

Mean

Conventional SD

2.58 3.19 0.08 9.24 42.20 3.10 6.17 0.13 41.38

0.70 1.04 0.01 3.37 12.14 0.29 0.66 0.01 6.49

2.74 3.43 0.08 9.35 42.79 3.03 5.80 0.14 39.03

0.75 1.28 0.01 3.69 14.79 0.29 0.64 0.01 7.03

2.69 3.56 0.08 9.65 43.43 2.94 5.91 0.13 40.79

0.82 1.29 0.01 3.77 18.34 0.40 0.85 0.02 9.46

2.89 4.02 ABC 0.08 10.15 AB 45.34 2.44 AC 4.84 AC 0.14 35.53 A

0.75 1.25 0.01 3.28 13.20 0.55 1.14 0.02 11.44

3.15 AC 4.11 ABC 0.08 10.28 AB 46.43 AB 2.52 AC 5.09 AC 0.14 37.26 A

0.73 1.19 0.01 3.33 13.52 0.42 0.87 0.01 7.27

* * * * * * *

Notes: * = significant main effect (p b 0.006). A = significantly different from barefoot, B = significantly different from Vibram Five Fingers, C = significantly different from Inov-8.

4. Discussion The current study aimed to determine whether running in BF and BFIS footwear caused different levels of knee and ankle loading patterns in comparison to conventional running trainers. This represents the first to investigate influence of BF and BFIS on PTF, PP and ATF in relation to conventional running footwear. In support of our hypothesis, the first key observation of the current investigation is that BF, Vibram Five Fingers and Inov-8 conditions were associated with significantly lower EM, PTF, PP and PTF load rate in relation to conventional footwear. This supports the findings of Bonacci et al. (2013a) who observed reduced eccentric quadriceps work during BF compared with conventional footwear. Bonacci et al. (2013b) hypothesized that this observation due to the reductions in stride length associated with barefoot running (Sinclair et al., 2013a). This positions the stance limb closer to the centre of mass, which reduces the moment arm of the quadriceps. This description may help clarify the mechanism by which increases in PTF and PP were observed when running in the conventional running footwear. This finding may also have clinical significance regarding the aetiology of injury as the widely accepted consensus regarding the development of patellofemoral pathology is that the symptoms are the function of excessive patellofemoral joint kinetics (Kulmala et al., 2013). Given the high incidence of patellofemoral disorders in runners (van Gent et al., 2007), running BF and in minimalist BFIS may be a potential mechanism by which runners area able to attenuate knee injury risk through reductions in knee joint loading. With respect to the ankle joint kinetics the findings suggest that conventional and Nike Free footwear were associated with reductions in MPF, ATF and ATF load rate in relation to BF and Inov-8. This observation may have clinical relevance for the pathogenesis of Achilles tendinopathy, a further pathology that is commonplace amongst recreational runners (Kulmala et al., 2013). The aetiology of Achilles tendinopathy is believed to be associated with repeated mechanical loading of the tendon. Repetitive tendon loads such as those experienced during distance running initiate collagen and extracellular matrix synthesis and tissue degradation. The results of the current investigation are in line with the observations of Bonacci et al. (2013a), suggesting significantly higher ankle plantarflexor moment during BF and BFIS. Reduced impact loading associated with BF is attributable to enhanced ankle joint energy absorption during the early stance phase (Lieberman et al., 2010). This increased plantarflexion contribution from the ankle joint may be the mechanism by which the reduced PTF and PP in BF and minimalist BFIS are accounted for. The knee joint function as an energy absorber during the stance phase is reduced in BF and BFIS, which results in reduced knee flexion excursion (Sinclair et al., 2013a, 2013b, 2013c, 2013d). The results of this investigation therefore suggest that this may be at the expense of greater loading of the Achilles tendon. As such, this study advocates that BF and minimalist BFIS may not be appropriate for runners who are predisposed to Achilles tendon pathology.

In addition the reduction in knee load associated should be contextualised taking into account the increased stride and step frequency, typically associated with BF and BFIS (Lieberman et al., 2010; Sinclair et al., 2013a, 2013b, 2013c, 2013d). As would be expected increased stride/step frequency is associated with reductions in joint energy absorption during the impact phase of running (Kulmala et al., 2013). Therefore whilst reductions in PTF and PP were noted per step during BF and BFIS, the amount of cumulative knee load may be minimally affected between BF, BFIS and conventional footwear, as the total number of footfalls required to achieve the same distance is greater. This is an interesting concept and there is currently a lack of epidemiological information regarding the effects of cumulative and magnitudal loads experienced by the musculoskeletal structures during gait with respect to the aetiology of chronic pathologies. The prevailing observation from the current investigation is that differences occurred chiefly between conventional/Nike Free running shoes in comparison to BF, whereas inconsistent deviations were found for the minimalist BFIS. It is likely that this finding relates to the greater divergence in midsole characteristics between conventional and Nike Free footwear compared to BF (Sinclair et al., 2013a), which results in the distinct running mechanics that lead to variations in knee and ankle kinetics (Sinclair et al., 2013a, 2013b). This also re-enforces the notion that BFIS footwear models are not analogous but most importantly shows that they may not necessarily replicate the characteristics of running BF. In conclusion, the observations of the current investigation show that running BF and in minimalist BFIS exhibit significant reductions in knee PTF and PP compared to running in conventional footwear. Given the proposed relationship between knee joint loading and patellofemoral pathology, the risk of the developing running-related knee injuries may be attenuated through BF and BFIS. However, taking into account the corresponding increase in ATF during BF and minimalist BFIS, this may in turn enhance the likelihood of chronic injuries to Achilles tendon. It is clear that additional analyses are required in order to provide prospective epidemiological analyses of BF and BFIS in relation to conventional running footwear and the influence of different knee and ankle load patterns on the aetiology of running injuries. Competing interests No conflict of interest will arise from any of the authors involved in this paper. Author contributions Both named authors have made a significant and substantial contribution to all aspects of the study. Each of the named authors provided a meaningful contribution to the conception, design, execution and interpretation of the study data in addition to writing, drafting and revising

Please cite this article as: Sinclair, J., Effects of barefoot and barefoot inspired footwear on knee and ankle loading during running, Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2014.02.004

J. Sinclair / Clinical Biomechanics xxx (2014) xxx–xxx

the paper itself. This paper is submitted with the agreement and approval of both authors. Funding No external funding was provided for this paper. Acknowledgements The authors would like to thank Glen Crook for his technical assistance during data collection and to wish him a very happy retirement. References Bonacci, J., Saunders, P., Hicks, A., Rantalainen, T., Vicenzino, B., Spratford, W., 2013a. Running in a minimalist and lightweight shoe is not the same as running barefoot: a biomechanical study. Br. J. Sports Med. 47, 387–392. Bonacci, J., Vicenzino, B., Spratford, W., Collins, P., 2013b. Take your shoes off to reduce patellofemoral joint stress during running. Br. J. Sports Med. http://dx.doi.org/ 10.1136/bjsports-2013-092160 (Epub ahead of print). Cappozzo, A., Catani, F., Leardini, A., Benedeti, M.G., Della, C.U., 1995. Position and orientation in space of bones during movement: anatomical frame definition and determination. Clin. Biomech. 10, 171–178. Ho, K.Y., Blanchette, M.G., Powers, C.M., 2012. The influence of heel height on patellofemoral joint kinetics during walking. Gait Posture 36, 271–275.

5

Heino, B.J., Powers, C.M., 2002. Patellofemoral stress during walking in persons with and without patellofemoral pain. Med. Sci. Sports Exerc. 34, 1582–1593. Kulmala, J.P., Avela, J., Pasanen, K., Parkkari, J., 2013. Forefoot strikers exhibit lower running-induced knee loading than rearfoot strikers. Med. Sci. Sports Exerc. 45, 2306–2313. Lieberman, D.E., Venkadesan, M., Werbel, W.A., Daoud, A.I., D'Andrea, S., Davis, I.S., Mang'eni, R.O., Pitsiladis, Y., 2010. Foot strike patterns and collision forces in habitually barefoot versus shod runners. Nature 463, 531–535. Powers, C.M., Lilley, J.C., Lee, T.Q., 1998. The effects of axial and multiplane loading of the extensor mechanism on the patellofemoral joint. Clin. Biomech. 13, 616–624. Self, B.P., Paine, D., 2001. Ankle biomechanics during four landing techniques. Med. Sci. Sports Exerc. 33, 1338–1344. Sinclair, J., Greenhalgh, A., Edmundson, C.J., Brooks, D., Hobbs, S.J., 2013a. The influence of barefoot and barefoot-inspired footwear on the kinetics and kinematics of running in comparison to conventional running shoes. Footwear Sci. 5, 45–53. Sinclair, J., Hobbs, S.J., Currigan, G., Taylor, P.J., 2013b. A comparison of several barefoot inspired footwear models in relation to barefoot and conventional and conventional running footwear. Comp. Exp. Phys. 9, 13–21. Sinclair, J., Hobbs, S.J., Taylor, P.J., Currigan, G., Greenhalgh, A., 2013c. The influence of different force measuring transducers on lower extremity kinematics. J. App. Biomech. (in press, http://www.ncbi.nlm.nih.gov/pubmed/23877002). Sinclair, J., Hobbs, S.J., Protheroe, L., Greenhalgh, A., 2013d. Determination of gait events using an externally mounted shank accelerometer. J. App. Biomech. 29, 118–122. van Eijden, T.M., Kouwenhoven, E., Verburg, J., Weijs, W.A., 1986. A mathematical model of the patellofemoral joint. J. Biomech. 19, 219–229. van Gent, R.N., Siem, D., van Middelkoop, M., van Os, A.G., Bierma-Zeinstra, S.M.A., Koes, B.W., 2007. Incidence and determinants of lower extremity running injuries in long distance runners: a systematic review. Br. J. Sports Med. 41, 469–480. Ward, S.R., Powers, C.M., 2004. The influence of patella alta on patellofemoral joint stress during normal and fast walking. Clin. Biomech. 19, 1040–1047.

Please cite this article as: Sinclair, J., Effects of barefoot and barefoot inspired footwear on knee and ankle loading during running, Clin. Biomech. (2014), http://dx.doi.org/10.1016/j.clinbiomech.2014.02.004

Effects of barefoot and barefoot inspired footwear on knee and ankle loading during running.

Recreational runners frequently suffer from chronic pathologies. The knee and ankle have been highlighted as common injury sites. Barefoot and barefoo...
497KB Sizes 0 Downloads 3 Views