Electrode sites were prepared for electromyography (EMG) by cleansing the skin with isopropylalcohol swabs, shaving body hair when necessary, and lightly abrading the skin surface with Medicotest disposable abraders before application of Medicotest N-50-E disposable electrodes. Two electrodes were applied over the belly of each muscle along the same direction as the muscle fibres and a reference was placed on a close bony prominence (e.g. clavicle; elbow). Electrodes from each muscle were attached to a four kilo-ohm preamplifier and all were taped to the skin surface to prevent any movement during the dynamic boxing task, taking care not to restrict the natural movement of the subject during boxing performance. The preamplifiers were connected to a MIE Medical Research Ltd. eight channel radiotelemetry transmitter with aerial. The differential preamplifiers had a gain of 4000, with a balance input of 10 mΩ, a common mode rejection ratio of 110 dB and a signal to noise ratio of -50 dB. The frequency pass band was 6 to 330 Hz (3 dB point).Prior to data collection each 20 minute self-selected warm-up comprising stretching, jogging and striking hand held coaching pads was performed. The height of the boxing dynamometer manikin, which was to be the target during each boxing trials, was matched to each subject's shoulder height from the floor. Following habituation, each subject was given 30 seconds to throw straight rear hand punches at maximal speed or force in turn to the head of a calibrated boxing dynamometer (Smith et al. 2000). Punches were thrown either singularly or in two/three punch combinations in a prescribed sequence from an audio cue (Akai, AJ- W248). Punch force data were sampled at 330 Hz and recorded using punch data acquisition software specified to run within Provec 5.0 software (MIE Medical research Ltd, Leeds, UK). Typically 19-20 punches were thrown by each hand during a punch sequence with movement recorded on a Panasonic video camcorder.Electromyography data was sampled for 15 seconds using Myodat 3.0 software during each data collection period and then stored to the computer hard disc for subsequent analysis. For each subject, six punches were chosen from the middle phase of the punch sequence and the main horizontal uniaxial component of impact force recorded for the qualitative muscular recruitment analysis. The integrity of the raw EMG for each selected punch was inspected, and if good then the linear enveloped EMG was processed. A sequence of muscle recruitment, based on the assumption of a linear recruitment model for the eight muscles, was determined for the maximal force and speed conditions for each subject and the group. The model utilised an increment of one for the position (within eight possible ones) a muscle was maximally recruited within a series of 1 to 8 for each subject.Relative activity levels of the rear hand right rectus femoris muscle in four punches delivered for maximal force and for those delivered for maximal speed were determined with related triaxial punch force data.