Sprinters will be recruited from the track and field team of National Taiwan Normal University (NTNU) in this study. After individuals' enrollments and baseline data collections, all subjects will receive all three different treatments (massage, cold water immersion and static stretching) in randomized orders a week apart, respectively. Outcome measures are: visual analogue scale (VAS) score, lower leg volume, pressure pain threshold and horizontal jump distance. All measurements will be recorded at baseline, immediately after exercise, immediately after treatment, and 10 minutes after treatment as the follow up.

One certified physical therapist will perform the massage regimens and techniques adapting from Swedish massage such as effleurage and petrissage that will be used in the massage group. Participants will first lay down in prone position, and the massage leg will be supported on the therapist's leg or on pillows with legs being positioned at approximately 30 degrees of knee flexion. This can promote venous return of the lower legs and also keep the ankle plantar flexors in a relative low muscle tension position. After positioning, long strokes will be applied from the distal to proximal part of gastrocnemius with approximately 20 repetitions and then 20 repetitions of circular strokes would be applied at the popliteal fossa. Finally, therapist will knead around participant's Achilles tendon for 10 repetitions.6 The combination of the techniques mentioned above will repeat 3 times in each subject, and the total duration of massage will be approximately 5 minutes.

Lower leg volume was measured with water displacement method using volumeter. First, the volumeter was filled with water until it overflowed into the pitcher under the spout. Before each measurement, the pitcher was emptied, dried and placed back under the spout. The height of the chair was adjusted to ensure that the lower leg was completed immersed into water (the level of the water reached the mark). The pitcher filled with water collected from the volumeter was then weighed so that the volume of the lower leg could be quantified. Since the density of water was close to 1, the weight of the outflowed water was equal to the volume of the lower leg. To ensure that the assessor could measure the same site at all time periods, the same mark was used in each testing session.

Pressure pain threshold was defined as the minimal amount of pressure that induced pain or discomfort. Participant was asked to lie in prone position with ankle sticking out of the edge of table to let the ankle slightly plantar flexed without resistance. The algometer was then placed at the muscle tendon junction of gastrocnemius on the testing leg. The assessor compressed the site with the algometer and gradually increasing pressure until the participant said "stop" whenever he or she started to feel pain or discomfort. At the same time, the second assessor read and recorded the value displayed on the algometer. Measurements were measured two times and the intervals between each measurement were around 5 seconds apart. These two readings were averaged and recorded for further data processing. The algometer showed good intra-rater reliability with an ICC of 0.96.