Literature search methodology
MEDLINE and ScienceDirect were searched for primary sources using six keywords: expertise, performance, technique, variability, swimming and start. These were pooled (via Boolean operation "OR") and combined (via Boolean operation "AND") with similarly pooled keywords related to swimming biomechanics. The proceedings of international congresses on biomechanics and swimming databases were also searched, from their earliest available records up to November 2012. Relevant articles were sought on Google Scholar, and the cited articles and reference lists of all included studies were carefully scrutinized. The articles analyzing swim-starts were restricted to those written in English. Full publications and abstracts were screened, and all relevant studies were retrieved. A standardized form was used to select the studies eligible for inclusion. Ultimately, 45 references and eight books were selected from the previously selected articles and books from the MEDLINE, ScienceDirect and Google Scholar searches; an additional 17 references were retrieved from the proceedings of sport sciences congresses. Disagreement was resolved by achieving consensus among the authors, who took into account the size of the population studied and the swimming skill level for inclusion.
The start in a swimming event
Recently, interest in swimming-specific research has begun to accelerate (Pelayo and Alberty, 2011). Indeed, Vilas-Boas (2010) noted that swimming is now one of the most investigated physical activities, based on the number of published research articles and the number of countries represented at international meetings. Part of this rise in interest may be related to the ongoing modifications in the swimming rules, driven by changes in swimming techniques and technologies, all of which have inspired new research directions. This includes the swim-start (SW 7 of the FINA rules), which has undergone several changes from a regulatory point of view. For example, on January 1, 2010, a new kick-start block was authorized, with a raised rear section to assist the track start technique (Omega OSB11). Competition analysis has provided information on the start time (to 15-m), turn times (7.5-m into and out of the wall), and finish time (5-m into the wall), as well as the stroke length, stroke rate and velocity, for each 25-m section of free swimming (Mason and Cossor, 2000). Moreover, the start time has been quantitatively evaluated in relation to the swimming, turn and finish times in order to assess its contribution to overall performance (Arellano et al., 1996; Costill et al., 1992; Lyttle and Benjanuvatra, 2005; Mills and Gehlsen, 1996; Vilas Boas et al., 2003). The results indicate that the start time to 15-m can account for anywhere between 0.8% and 26.1% of the total race time, depending on the event (Lyttle and Benjanuvatra, 2005) (i.e., the latter percentage reflecting the percentage in sprint events). Moreover, contrary to the block starts in long-distance events, in which the athlete has to accelerate from zero to full running speed, dive swim-starts enable swimmers to enter into the water faster than average swimming speed, which further underlines the great importance of the swim-start in sprints. Effective diving techniques enable swimmers to exploit the speed generated during the dive and are in line with the principle of efficiency that drives every phase of the competitive event (Kilduff et al., 2011; Lyttle and Blanksby, 2011).
Analysis of swimming start kinematics Methodology
The studies on the swim-start have analyzed several parameters. Kinematic analyses of swim-start behavior and performance, for example, have usually compartmentalized the start into distinct phases, such as block time, flight time and underwater time (Arellano et al., 1996; Cossor and Mason, 2001; Vilas-Boas et al., 2003). More recent studies have assumed that the start actually begins with the reaction to the start signal and the push from the block (Benjanuvatra et al., 2007; Bishop et al., 2009; De la Fuentes et al., 2003; Slawson et al., 2013). These trials were recorded at 50 Hz with a digital video camera placed perpendicularly to the direction of movement. Vantorre et al. (2010a) used both fixed cameras (placed at 5-m and 15-m) to determine phase limits and underwater mobile cameras on a trolley to analyze qualitative variables and stroking parameters like stroke length or frequency. The forces applied during the push from the starting block were analyzed via custom-built, instrumented starting blocks. Force curves measured the impulse in the horizontal and vertical axes (in N-kg-1) (Benjanuvatra et al., 2007; Blanksby et al., 2002; Lee et al., 2001; Slawson et al., 2013; Vantorre et al., 2010b, 2010c; Vilas-Boas et al., 2003; West et al., 2011). The kinetic analysis of the block phase quantified the impulse and described its direction relative to the direction of movement (Benjanuvatra et al., 2007; Blanksby et al., 2002; Lee et al., 2001; Slawson et al., 2013; Vantorre et al., 2010b, 2010c; Vilas-Boas et al., 2003).
Several studies of swim-start phase kinetics, particularly the reaction time on the starting block and the flight and entry phases, have drawn parallels with the start in track and field (Ayalon et al., 1975; De la Fuentes et al., 2003; Issurin and Verbitsky, 2003; Kruger et al. 2003; Miller et al., 2003; Vilas-Boas et al., 2003; Zatsiorsky et al., 1979). However, from a biomechanical point of view, these starts differ in many ways. Moreover, among swimmers, the starts also differ according to specialty. Sprint swimmers need to rotate backwards to bring themselves upright, whereas longer-distance swimmers need to focus on the distance covered while in the air and the body orientation at water entry. Here, breaking down the swim-start is not only a spatial matter, but also a matter of motor changes during the overall start movement. From this perspective, studies on the block phase (Benjanuvatra et al., 2007; Vantorre et al., 2010a) have shown that two distinct actions must be optimized: a rapid reaction to the start signal and high impulse generated over the starting block. The studies on the block phase have usually been kinetic analyses focused on the force applied to the block or on training programs designed to improve the start (Bishop et al., 2009; Breed and Young, 2003; De la Fuentes et al., 2003; Lee et al., 2001). The reaction time needs to be as brief as possible, while the movement phases on the block need to last long enough to maximize the swimmer's impulse to achieve high horizontal velocity (Breed and Young, 2003). In other words, a compromise needs to be struck between spending too much time on the block to create more force and spending too little time on the block to minimize the time deficit and avoid being "left at the start" (Lyttle et al., 1999).
Flight and entry phases
Breaking down a swim-start into its component parts can be challenging as the phases are not always clear cut. Maglischo (2003) defined water entry as the moment when the hand enters the water. This definition is widely used to determine the end of the flight phase, during which swimmers need to jump as far as possible and travel the maximum distance at the high velocity developed during the block phase (Hubert et al., 2006; Sanders and Byatt-Smith, 2001). Ruschel et al. (2007) reported that flight duration is not correlated with start time but that flight distance is one of the variables that determine starting performance (r = -0.482). Maglischo (2003) noted that the block phase strongly influences the flight phase by imposing a compromise between the pike and flat styles for the aerial trajectory (Maglischo, 2003). The pike start has a longer start time, greater take-off and entry angles, and a shorter distance to head entry into the water than the flat start (Counsilman et al., 1988). Wilson and Marino (1983) showed a shorter 10-m start time, greater entry angle, shorter distance to water entry, and greater hip angle at entry for the pike start than for the flat start. However, after five training sessions, Kirner et al. (1989) reported that the grab start/flat entry showed a shorter 8-m start time and a smaller entry angle than the grab start/pike entry. Thus, the flat start aims for a quick entry into the water using a flatter body position and earlier stroking. The pike start creates a smaller hole for water entry (i.e., angle of entry more vertical to the water surface) with higher velocity due to the influence of gravity, but it requires a horizontal (body position from the surface) then vertical(until break out the water surface) underwater recovery, which causes higher resistance. Vantorre et al. (2010a) studied swim-starts and found that strategies differ even among elite swimmers. These authors observed that the swim-start profiles included differences in how the limbs were used to achieve specific trajectory styles, such as the Volkov start, with the arms back during the leg impulse, or the flight style start, with the arms directly in front of the head (Vantorre et al., 2010a). However, the swimmer's task during the flight phase is not merely to go as far as possible. Mclean et al. (2000) and Vantorre et al. (2010a; 2010b) showed that swimmers must also generate enough angular momentum to make a clean entry into the water, which means that they need sufficient time to rotate while in flight in order to enter the water through a small hole. Arm movements influence angular momentum and during the forward rotations of the swim-start, a forward arm swing decreases rotation and, inversely, backward rotations increase body rotation (Bartlett, 2007a). Therefore, to manage the angular momentum generated during the block phase, swimmers can make a flat start (less angular momentum and a flat trajectory) or a Volkov start with a backward arm swing (more angular momentum and...