So, all the temporal variables and the PP from the frequency variables show differences between genders. Considering the performance level, most the differences are not significant (p<0.05). However, in the gender and performance analysis there are only differences between male levels in RMS and PP (p<0.05). Also it is confirmed the existence of two types of frequency spectrums in the front crawl swimming. About the types of frequency spectrums of the acceleration The observation of the spectrum profiles allowed us to concentrate the swimmers into different groups, regarding the number of peaks shown, similar to that of Tella et al. (2008) and Madera et al. (2010) observed in their respective works.

In this study and with the intention of differentiating those spectrums that visually presented smaller variation (type 1), those spectrums with more than one relevant PP regarding the rest of the signal and those with a wide range of frequencies in the PP were considered as type 2. So the type 1 spectrums show only one PP associated to only one frequency. This work confirms that there are different types of frequency spectrums of the acceleration in front crawl swimming. Type 1 and 2 spectrums are common in both genders and performance levels. Regarding the percentage of type 1 spectrums in this study (43.1%) it should be noted the difference with the 27.85%, observed by Madera et al. (2010). The possible cause may be the two instead of three groups that were utilized in that work. The analysis of the spectrum profiles have been done by several authors with the aim of explaining them.

Holm��r(1979) identified the front crawl spectrum with a low SF (PPF<1Hz) and suggested that it represented the propulsive actions of the arms. Furthermore, this author identified several smaller peaks (between 1.5 and 10 Hz) and suggested that these may correspond to the movement of the legs. As it has been used in previous studies, the methodology in this work has reduced the registered accelerations (Madera et al., 2010: 1 to 10Hz; Tella et al., 2008: 1 to 20Hz) to frequencies lower than 1Hz by filtering the signal with a band-pass filter of 1�C10Hz, after considering that the accelerations below 1Hz would represent the global actions of a complete cycle (i.e. bodyroll). Also the accelerations with frequency higher than 10Hz have been reduced, because the spectrums that were found by Tella et al.

(2008) did not show important peaks in those frequencies. Then to preserve only those frequencies of interest for the study, the accelerations whose frequencies were lower Drug_discovery than 1Hz and were higher than 10Hz were reduced to achieve a representation of only the ones that could be related to intra-cycle accelerations. Thus the analyzed spectrums may represent the propulsive actions in front crawl swimming (i.e. arms and legs) with two different types that exist at both genders and performance levels.