Since the appearance of the skating techniques in cross-country skiing, a number of studies have made comparisons of the physiological demands (2,9,12-14,16-18,23) and kinematics (1,3,10,26-28) of different techniques. However, the available information on forces generated with different skiing techniques is quite limited. The studies of classical (6,7,19,20,22) and skating (24,25,29) techniques have examined small numbers of subjects, and have generally only included representative data with little analysis. Additionally, no study has made kinetic comparisons among different techniques with the same subjects.
Based upon preliminary data, it has been reported that the majority of propulsive forces are generated through the poles with ski skating (24). Because the propulsive forces from the upper body seem particularly important in these techniques, an examination of poling forces with different skating techniques is warranted.
This paper reports new information about poling forces during ski skating. Poling forces during roller skiing on a slight grade were compared among three different ski skating techniques and across a wide range of speeds. The double poling technique, which is common to both skating and classical cross-country skiing, was also included in this analysis.
Subjects. Nine adult male cross-country skiers completed the study. Each subject was highly skilled in cross-country skiing as attested by a top 200 placing in a Worldloppet race (American Birkebeiner, U.S. or Transjurassienne, France) during the 3-yr preceding the study. All subjects specialized in the skating techniques, reporting that they spent 93 ± 14% (mean ± SD) of their training time using these techniques. Other selected characteristics of the subjects are outlined in Table 1. Before participating, each subject received an explanation of the procedures and gave written informed consent in accordance with our institution and the American College of Sports Medicine.
Skiing techniques. In this study, three different skating techniques were examined. Because the terminology of the skating techniques is not standardized, the skating techniques examined in the present study will be briefly described. The V1 skate (V1; also referred to as the offset technique) and V2-alternate techniques (V2A; also referred to as the 2-skate, Gunde skate, and open field skate) use a double pole plant as weight is transferred to one ski. The side of the body supporting the weight during all or most of the poling phase is considered the strong side, with the opposite side being referred to as the weak side. Pole plants are considered to be asymmetrical and slightly asynchronous for V1, and symmetrical for V2A (1). The V2 skate technique (V2; also referred to as the 1-skate) has a symmetrical double pole plant during weight transfer to each ski (i.e., two pole plants for each full cycle). Also included in the present study was the double pole technique (DP). With this technique, all the propulsive forces are applied through symmetrical and synchronous pole plants.
Protocol and measurements. The roller skiing tests were performed on a straight and smooth asphalt roadway. The test zone was a 70-m section with constant grade of 2.1% as determined from measurements with an automatic level and transit (Model AL6-18 David White Instruments, Menomonee Falls, WI). An adequate distance of comparable grade preceding the test zone was used to allow the skiers to reach the desired speed before any measurements were performed. The surface was dry and free of debris during testing.
The various techniques and speeds that are examined in this paper are displayed in Table 2. For a given subject, all bouts of roller skiing were performed on the same day in association with the tests described in the accompanying paper (21). The order in which the subjects performed the different skiing techniques was randomized, with the order of speeds always progressing from slowest to fastest for each technique. For the submaximal speeds, the skiers were paced by an investigator riding approximately 5 m in front of the skier on a bicycle equipped with a calibrated electronic speedometer. Each bout was timed independently across the test zone to determine the actual speed. If the measured speed was not within 0.5 km·h−1 of the desired speed, the bout was repeated. Actual speeds for each condition are shown in Table 2. Immediately after each bout of roller skiing, a rating of perceived exertion (RPE) using the Borg 6-20 scale (4) was requested. The subjects rested a minimum of 2 min between bouts. With V1 and V2A, the subjects were instructed to consistently use the same strong side across all speeds.
The same pair of roller skis (V2-820, Jenex, Inc., Amherst, NH) was used for all tests. Rolling resistance was measured before the study and after completion of all testing using techniques described previously (11,15,16). No measurable change in rolling resistance occurred over the duration of the study. The coefficient of friction (μ) for rolling was found to be dependent on the velocity (V in km·h−1) and was described by the linear regression μ = 0.018 + 0.00042·V (r2 = 0.88; P < 0.01). Thus, across the speeds used in this study, μ ranged from 0.023 to 0.027.
Air temperature and relative humidity during testing varied from 2.3 to 19.5°C and 45 to 58%, respectively. Wind velocity during each bout was measured with an anemometer (Dwyer Instruments, Michigan City, IN) and ranged from 0 to 0.7 m·s−1.
Pole force data were collected during each bout of roller skiing. Aluminum ski poles (Rollerlite, Swix, Lillehammer, Norway) were modified so that single-axial force transducers could be mounted 1 cm below standard handles. Piezoelectric force transducers (model 208B03, PCB Piezotronics, Inc., Depew, NY) were selected for their ability to reject cross-talk due to bending moments (29). Each force transducer weighed 23 g. Data from the transducers were recorded by a portable data logger (Tattletale model 2B, Onset Computer Corp., N. Falmouth, MA) carried by the subject in a waist-bag (total mass 1.0 kg). Data collection was triggered telemetrically by an investigator when the subject reached the beginning of the test zone. Data were recorded for 15 s at a sampling frequency of 100 Hz. After each bout of roller skiing, the data were downloaded to a computer for subsequent analysis. Calibration of the force transducers was performed in the field before the testing of each subject.
Data obtained from the load cells were first smoothed by a three-point moving time average. The force curves were then reviewed, and consecutive cycles that appeared representative of the condition were selected for analysis. Five consecutive cycles were analyzed for V1, V2A, and DP, and three consecutive cycles (i.e., six poling cycles) were analyzed for V2. For each analyzed cycle, the beginning of pole plant was determined as the first positive difference between two consecutive samples greater or equal to 0.3 V (about 27 N). The end of the pole plant was determined as the first sample point below 0 V. One cycle was defined as the time between consecutive right pole plants for V1, V2A, and DP and the time between every other right pole plant for V2. Cycle time (CT), poling time (PT), poling recovery time (RT = CT − PT), poling duty cycle (DC = PT·CT−1), peak poling force (PF), impulse (integration of the force-time curve), average poling force (AF = impulse·PT−1), and average poling force across the entire cycle (ACF = impulse·CT−1) were determined. With V2, PT was considered as the average of the two pole plants within the cycle, RT was calculated as (CT·2−1)-PT, and DC was calculated as 2·PT·CT−1. The time-related variables are displayed in Figure 1.
Six different ski pole lengths, ranging from 140.0 to 167.5 cm, were available for the subjects to use. The mean (±SD) pole length chosen by the subjects was 162.3 ± 2.8 cm (89.5 ± 2.3% of body height). Poles tips were sharpened so that good purchase was maintained during all testing.
Statistical analysis. Mean values for both poles across the examined cycles of each study variable were compared among techniques and across speeds with two-way (technique × speed) ANOVA with repeated measures. Comparisons of poling forces and times were also made between strong- and weak-side poles with V1 and V2A using two-way (side × speed) ANOVA with repeated measures. When F-values were significant, individual comparisons were made with the Newman-Keuls post hoc test. Differences between the start time of pole plants for the strong- and weak-side poles with V1 and V2A were analyzed with t-tests. For all analyses, a P-value of 0.05 was accepted as the level of statistical significance.
Maximal speed and RPE. Maximal velocities with V1, V2, and V2A were not significantly different but DP averaged 9% slower (P < 0.05) than the skating techniques (see Table 2). RPE values rose with speed (P < 0.0001) but were not statistically different among techniques (Fig. 2).
Forces. PF, AF, and ACF increased (P ≤ 0.05) with speed for each technique except V2, which showed no significant change in ACF with speed (Fig. 3). Significant main effect differences between techniques were present with post hoc analyses revealing that PF, AF, and ACF values were higher (P < 0.01) with DP than the skating techniques. PF and ACF values were higher (P < 0.01) with V2 than V1 and V2A, and ACF was also higher (P < 0.05) with V1 than V2A. AF was not different among the skating techniques. Significant (P < 0.05) technique by speed interaction effects were present with post hoc analyses revealing that at maximal speed, PF was not different between DP and V2, and ACF was not different between V1 and V2.
Comparing the strong- and weak-side poles, PF, AF, and ACF were similar between sides with V2A (Fig. 4). However, with V1, the values for the strong-side pole were higher than the weak-side pole by 24.2% (P = 0.006), 12.6% (P = 0.005), and 18.1% (P = 0.02) for PF, AF, and ACF, respectively (Fig. 5). For the latter technique, the lack of a significant interaction effect of pole side with speed indicates that the difference between the poles remained similar across speeds.
Timing. PT and RT decreased (P < 0.01) with speed for all techniques (Fig. 6). Significant main effect differences between techniques were present with post hoc analyses revealing that PT was longer (P < 0.01) with DP compared with V2 and V1 and longer (P < 0.01) with V2A compared with V2 and V1. RT was longer (P < 0.01) with V2A compared with the other techniques and longer (P < 0.01) with V1 compared with V2 and DP. RT was not significantly different between DP and V2. DC was significantly different (P < 0.01) among all techniques, with a rank order of DP > V2 > V1 > V2A. DC also showed a significant (P = 0.0001) main effect of speed and a significant (P < 0.0001) technique by speed interaction. Post hoc analysis revealed that the DC values at the lowest submaximal speed were greater than (P < 0.01) the other speeds with the skating techniques, whereas the DC values at all of the submaximal speeds were less than (P < 0.05) at maximal speed with DP.
The time at which poling began for the strong- and weak-side poles was significantly different (P = 0.0001) with V1 (mean ± SD difference = 31 ± 24 ms), with the beginning of compression on the weak-side pole beginning sooner. In contrast, there was no significant difference in the time at which poling began for the strong- and weak-side poles with V2A (mean ± SD difference = 4 ± 20 ms). PT was not significantly different between the strong and the weak pole with V1 or V2A.
A significant main effect (P < 0.0001) of speed on cycle rate (CR) was present (Fig. 7). CR was also significantly different (P < 0.01) among all techniques with a rank order of DP > V1 > V2A > V2. A significant (P < 0.0001) technique by speed interaction effect was present, and inspection of the data suggests that the increase in CR with speed was greater with DP than the skating techniques.
A significant main effect (P < 0.0001) of speed on cycle length (CL) was observed, and differences (P < 0.01) were present among all techniques following the rank order of DP < V1 < V2A < V2 (Fig. 7). A significant (P < 0.0001) technique by speed interaction effect was also present for CL. Post hoc analysis revealed that the evolution in CL across the submaximal speeds was similar among the skating techniques, with CL showing statistically significant increases. However, CL did not change significantly across submaximal speeds with DP. Statistically significant decreases (P < 0.05) in CL were present between the highest submaximal speed and maximal speed with V2 and DP.
This study systematically examined poling forces for different skating techniques and the double pole technique. Examination of poling forces provided useful information about the magnitude of force production as well as a number of timing-related variables. The most important findings of this study were that 1) poling forces with the skating techniques were considerably lower than the values previously reported, 2) the use of the upper body was greater with V2 than the other skating techniques whereas the poling forces with V2 did not appear to be as effectively applied as compared with V2A, and 3) forces applied through the poles appeared to contribute relatively less to the additional propulsive forces required to increase speed with V2 compared with V1, V2A, and DP. Because studies have not been performed to compare poling forces and kinematics during roller skiing and on-snow skiing, the application of the present results to on-snow skiing conditions is not clear.
Maximal speed and RPE. Although the duration of most competitions in cross-country skiing is much too long for skiers to sustain maximal sprinting speeds, there are certainly conditions when maximal speed is an important factor in racing. For example, sprinting speed at the beginning and end of races may be of considerable importance in determining race results. The present study demonstrated no differences in maximal speeds among the three skating techniques. This finding is consistent with a previous study of on-snow skiing that demonstrated that V1, V2, and V2A have similar maximal speeds at 0 and 6% grades (5). Interestingly, not all the subjects in our study achieved their highest velocity with the same technique, indicating that skiers might have individual preferences about which technique to use in a sprint.
The present study also demonstrates that the skating techniques were faster than DP. In a previous study (10), there was no significant difference in maximal speed between DP and V1 during roller skiing on a flat track. It is likely that the difference in maximal speed was evident in the present study because of the greater resistive forces acting on the skier. Higher resistive forces in the present study resulted from a higher roller ski rolling resistance (nearly 0.03 at maximal speeds compared with 0.02 in the previous study) and the roller skiing being performed on a slight uphill. Furthermore, a longer sprinting distance was used in the present study. Subjects were required to maintain maximal speed over a 70 m distance compared with 20 m in the earlier study.
Previous studies have demonstrated that the perceived demands at submaximal speeds are higher with DP than V1 during roller skiing (13) and on-snow skiing (9) on flat terrain under steady-state conditions. In the present study, mean RPE values tended to be higher for DP compared with V1, but statistical significance was not reached. The lack of a statistical difference may be at least partially because the bouts of roller skiing in the present study were relatively short. RPE values were also not significantly different among the three skating techniques. These results are consistent with observations that skiers use all three skating techniques on slight grades during competitions. Our findings of similarity in perceived exertion among the skating techniques also corroborates the findings of Bilodeau and colleagues (2) that velocities at competition effort were similar among skating techniques over various terrain.
Forces. The present study demonstrated that the highest PF values among the skating techniques were with V2 which ranged from 22.7 ± 2.3% to 31.5 ± 5.6% BW across the speeds that were tested. These PF values were significantly lower than the values of 26.4 ± 3.4% to 32.6 ± 3.3% BW that were observed for DP. The PF values we observed with ski skating were also considerably lower than the values of 50-60% BW reported by Smith (24) but considerably higher than the values of 13-17% BW reported by Pierce and coworkers (22) for diagonal striding. In comparing data with other studies, it should be recognized that poling forces are affected by skiing speed and the magnitude of slope and other resistive forces. Data were collected at slopes of 9 and 14% and velocities up to sprinting speed by Smith (24), but slope and speeds were not reported by Pierce et al. (22). Nevertheless, it would appear that peak poling forces required for ski skating may not necessarily be 2-4 times those with diagonal striding as has been previously suggested (24). Using the diagonal stride data of Pierce et al. (22), it appears that peak poling forces with ski skating may be 1.5-2 times those with diagonal striding.
ACF was significantly higher with V2 compared with V1 and V2A. This finding is at least partially because each complete cycle with V2 includes two pole plants with each pole rather than one. However, given that PF was also significantly higher across velocities with V2 compared with V1 and V2A, it must be concluded that upper body demands with V2 are greater than with the other two skating techniques studied.
ACF increased with velocity for DP, V1, and V2A but did not change with velocity for V2. This finding suggests that V2 relies more on the lower body for generation of the additional propulsive forces to increase velocity. A consideration is that poling forces might be relatively more effectively applied with V2 as speed increases. However, we do not believe this is the case. From our observations, it does not appear that poling forces with V2 compared with the other skating techniques become better directed along the line of travel as speed is increased. Moreover, PT decreased similarly with increasing speed for V2 as with the other techniques, suggesting that the final phase of the poling motion was not prolonged with V2 at the higher speeds. The last phase of poling is when the component of force directed along the line of skier travel is the greatest because the angle of the pole with the ground decreases during the poling motion.
At maximal speed, PF, AF, and ACF were found to be lower with V2A and V1 than DP. This finding suggests that the upper body is not fully taxed when skating at maximal speed with these two skating techniques. The comparison of V2 with DP at maximal speed revealed lower AF and ACF values for V2 but no statistically significant difference between techniques for PF. Although these results are not definitive, it appears that the upper body is also not fully taxed at maximal speed with V2.
Based on visual observations, poling has been described as asymmetrical and asynchronous with V1 and symmetrical with V2A (1). The present data demonstrated there were no differences between V1 and V2A in the combined poling force variables from the two poles. However, comparison of the strong- and weak-side poles during V1 revealed statistically significant differences with the strong-side pole producing 24.2%, 12.6%, and 18.1% higher values for PF, AF, and ACF, respectively. There was also a small (31 ± 24 ms) but significant difference in the time at which poling began between strong- and weak-side poles with V1, although poling times were not significantly different between the strong and weak poles. In contrast, for V2A there were no differences between poles in forces or the time at which pole plant occurred. Thus, the present findings confirm earlier impressions from visual observations.
Timing. In the present study, PT was found to be longer with V2A than V1 and V2. Because the pole plant position was probably comparable between V2A and V2, it is likely that the final phase of the poling motion was abbreviated with V2. As described above, the last phase of the poling motion is when the component of force directed along the line of travel is greatest because the angle of the pole with the ground decreases during the poling motion. Thus, even though the demands on the upper body are greater with V2 than the other skating techniques, it appears that the poling forces are less effectively applied with V2 compared with V2A. Future studies combining kinetic and kinematic analyses will be required for verification that this is the case.
It should be considered that poling forces may serve a purpose other than generation of propulsive forces. For instance, the poles may be used to assist in maintaining balance. The shorter RT values with V2 observed in the present study might be linked with balance limitations. Thus, further investigations should be performed with elite skiers, who might be less likely to have limitations in balance, to confirm the suggestion from the present study that poling effectiveness is lower with V2.
DP differs from other cross-country skiing techniques in that all of the propulsive forces are generated through the poles. The present findings suggest that the propulsive forces with DP are generated through reliance on a longer PT compared with V1 and V2 and a shorter poling RT compared with V1 and V2A, resulting in a greater poling DC compared with all of the skating techniques studied. The previous discussion relative to poling effectiveness also applies to the situation with DP. The longer PT with DP compared with V1 and V2 probably means that the final phase of the poling motion is extended with DP so that poling effectiveness is enhanced. Also of interest to note was the finding that poling DC increased across speeds for DP rather than decreasing across speeds as with the skating techniques. This means that the decrease in RT occurred to a greater extent than the decrease in PT as speed was increased with DP.
In a kinematic observation of V2A during a flat section of the men's 50-km competition of the 1992 Winter Olympic Games (27), it was found that skiers had mean PT values of 0.31 s at 22.7 km·h−1. In another study of on-snow skiing on flat terrain, PT values with V2A were found to be 0.27 s at 20.8 km·h−1(1). At comparable speeds, the PT values for V2A observed in the present study were similar to those reported in these previous studies. However, DC was higher (approximately 27% compared with 18-22%) in the present investigation compared with these on-snow studies (1,27). Further comparison of results show that DC in the present study for V1 and V2 was higher than the study of Bilodeau et al. (1), with mean values being 30.4% vs 22.4% and 34.2% vs 29.3% for V1 and V2, respectively. Differences between roller skiing and on-snow skiing may account for the variation in findings in that the angle of pole plant and the change in pole angle as the cycle progresses may differ between roller skiing and on-snow skiing. However, the most obvious explanation for the higher DC observed in our study is that the resistive forces acting on our skiers were higher. The ski on snow coefficient of friction is dependent on snow and humidity conditions, and has been reported to vary from 0.02 to 0.07 (8,23,30). Unfortunately, values for the coefficient of friction of the ski on snow were not defined in the previous studies, so the influence of this resistive force cannot be evaluated. The resistive force to overcoming gravity was higher in the present study since it was performed at a slight grade rather than on flat ground.
CR increased with speed for all techniques examined in the present study. CL increased across the submaximal speeds and then showed a plateau or decrease at maximal velocity with the skating techniques but showed no increase across submaximal velocities and then decreased at maximal velocity with DP. Furthermore, CR was higher and CL was lower with DP compared with the skating techniques. These findings demonstrate the different mechanism utilized for control of speed with DP compared with the skating techniques. When propulsive forces are only generated through the poles, RT is limited to the extent that CL is not increased across submaximal velocities. However, generation of propulsive forces through the skis allows propulsive forces to be applied over a larger proportion of the cycle. As a consequence, CL can be longer and can increase even as CR is increased.
Among the skating techniques, CR followed the rank order of V1 > V2A > V2. CR was lower for V2A than V1 because PT and RT values were greatest with V2A. V2 had the lowest CR, even though this technique had the shortest RT and the same or shorter PT compared with the other skating techniques, because two pole plants are used for each complete cycle with V2. The present finding of rank order for CR (and the opposite for CL) is in agreement with the results of Bilodeau et al. (1) who examined on-snow skiing at 80% of maximal speed and Boulay et al. (5) who examined on-snow skiing at maximal speed. Our results extend these earlier findings to demonstrate that this relationship among skating techniques holds across a wide range of speeds.
In summary, the skating techniques examined in the present study were found to be similar with regard to maximal speed, the RPE values at given speeds, and the manner in which CL and CR are adjusted to change velocity. Nevertheless, the present results indicate that with V2, the use of the upper body is greater than with the other skating techniques although there appears to be a relatively greater reliance on the lower body for generation of the additional propulsive forces required to increase velocity. It appears that poling forces are not as effectively applied with V2 as with V2A, but combined kinetic and kinematic analyses must be conducted for confirmation. Additionally, V1 differs from V2A in that the poling action is asymmetrical and asynchronous with V1, and the CR is higher with V1 because of the shorter poling and upper body recovery phases. In contrast to the skating techniques in which propulsive forces are generated through the skis as well as the poles, all propulsive forces with DP are applied through the poles, resulting in greater poling forces. In addition, increases in velocity with DP are achieved through increases in CR to the extent that CL is not increased even across submaximal velocities. For all techniques, PT and RT decrease with increasing speeds, but for DP the effect on RT is more pronounced, causing an increase in DC rather than the decrease observed for the skating techniques.
1. Bilodeau, B., M. R. Boulay, and B. Roy. Propulsive and gliding phases in four cross-country skiing
techniques. Med. Sci. Sports Exerc.
2. Bilodeau, B., B. Roy, and M. R. Boulay. A comparison of three skating techniques and the diagonal stride on heart rate responses and speed in cross-country skiing
. Int. J. Sports Med.
3. Bilodeau, B., K. W. Rundell, B. Roy, and M. R. Boulay. Kinematics of cross-country ski racing. Med. Sci. Sports Exerc.
4. Borg, G. Perceived exertion as an indicator of somatic stress. Scand. J. Rehabil. Med.
5. Boulay, M. R., K. W. Rundell, and D. L. King. Effect of slope variation and skating technique on speed in cross-country skiing
. Med. Sci. Sports Exerc.
6. Dal Monte, A., S. Fucci, L. M. Leonardi, and V. Trozzi. An evaluation of the diagonal stride technique in cross-country skiing
. Int. J. Sport Biomech.
7. Ekstrom, H. Force interplay in cross-country skiing
. Scand. J. Sports Sci.
8. Frederick, E. C. Mechanical constraints on Nordic ski performance. Med. Sci. Sports Exerc.
9. Hoffman, M. D., and P. S. Clifford. Physiological responses to different cross country skiing techniques on level terrain. Med. Sci. Sports Exerc.
10. Hoffman, M. D., P. S. Clifford, and F. Bender. Effect of velocity on cycle rate
and length for three roller skiing techniques. J. Appl. Biomech.
11. Hoffman, M. D., P. S. Clifford, B. Bota, M. Mandli, and G. M. Jones. Influence of body mass on energy cost of roller skiing. Int. J. Sport Biomech.
12. Hoffman, M. D., P. S. Clifford, P. J. Foley, and A. G. Brice. Physiological responses to different roller skiing techniques. Med. Sci. Sports Exerc.
13. Hoffman, M. D., P. S. Clifford, G. M. Jones, B. Bota, and M. Mandli. Effects of technique and pole grip on physiological demands of roller skiing on level terrain. Int. J. Sports Med.
14. Hoffman, M. D., P. S. Clifford, P. B. Watts, et al. Physiological comparison of uphill roller skiing: diagonal stride versus double pole. Med. Sci. Sports Exerc.
15. Hoffman, M. D., P. S. Clifford, P. B. Watts, K. P. O'Hagan, and S. W. Mittelstadt. Delta efficiency of roller skiing with the double pole and diagonal stride techniques. Can. J. Appl. Physiol.
16. Hoffman, M. D., G. M. Jones, B. Bota, M. Mandli, and P. S. Clifford. In-line skating: physiological responses and comparison with roller skiing. Int. J. Sports Med.
17. Karvonen, J., R. Kubica, S. Kalli, B. Wilk, and S. Krasicki. Effects of skating and diagonal techniques on skiing load and results in cross-country skiing
. J. Sports Med. Physiol. Fitness
18. Karvonen, J., R. Kubica, B. Wilk, J. Wnorowski, S. Krasickiand, and S. Kalli. Effects of skating and diagonal skiing techniques on results and some physiological variables. Can. J. Sport Sci.
19. Komi, P. V. Forces measurements during cross-country skiing
. Int. J. Sport Biomech.
20. Komi, P. V., and R. W. Norman. Preloading of the thrust phase in cross-country skiing
. Int. J. Sports Med.
21. Millet, G. Y., M. D. Hoffman, R. B. Candau, and P. S. Clifford. Poling forces during roller skiing: effects of grade. Med. Sci. Sports Exerc.
22. Pierce, J. C., M. H. Pope, P. Renstrom, R. J. Johnson, J. Dufek, and C. Dillman. Force measurement in cross-country skiing
. Int. J. Sport Biomech.
23. Saibene, F., G. Cortili, G. Roi, and A. Colombini. The energy cost of cross-country skiing
and the effect of the friction of the ski. Eur. J. Appl. Physiol.
24. Smith, G. A. Biomechanics of cross-country skiing
. Sport Med.
25. Smith, G. A. Biomechanical analysis of cross-country skiing
techniques. Med. Sci. Sports Exerc.
26. Smith, G. A., J. B. Fewster, and S. M. Braudt. Double poling kinematics and performance in cross-country skiing
. J. Appl. Biomech.
27. Smith, G. A., and B. S. Heagy. Kinematic analysis of skating technique of Olympic skiers in the men's 50-km race. J. Appl. Biomech.
28. Smith, G. A., J. McNitt-Gray, and R. C. Nelson. Kinematic analysis of alternate stride skating in cross-country skiing
. Int. J. Sport Biomech.
29. Street, G. M., and E. C. Frederick. Measurements of skier-generated forces during roller-skiing skating. J. Appl. Biomech.
30. Street, G. M., and R. W. Gregory. Relationship between glide speed and Olympic cross-country ski performance. J. Appl. Biomech.