Document Sample

Implications for Improved Technique and Safety

Peter D. Milburn

Department of Biomedical Science

The University of Wollongong

Wollongong, NSW



Summary ..........................................................    1

1     Introduction ...............................................    3

      1.1 Scrum-related Injury ...................................    4

             1.1.1 Incidence of Serious Spinal Injury .............       4

             1.1.2 Mechanisms of Spinal Injury ....................       7

2     Analysis of Contemporary Scrummaging Technique .............    8

      2.1 Engagement .............................................    9

      2.2 Sustained Scrummaging .................................. 13

      2.3 Binding Techniques ..................................... 16

      2.4 Player Contribution .................................... 17

      2.5 Performance Level ...................................... 21

      2.6 Positional Changes ..................................... 22

3.    Conclusions ................................................ 24


      In the game of rugby union, the scrum epitomises the physical nature

of the game. It is both a powerful offensive skill, affording a base for
attacking play, and a defensive skill in denying the opposition clean

possession.     However,      the    scrum    has      also    been   implicated     in   a    large

proportion of serious spinal injuries in rugby union.                             The majority of

injuries     were    found    to    have    occurred     at    engagement    where    the     forces

experienced by front-row players (more than two-thirds of a tonne shared

across the front-row) can exceed the structural limits of the cervical

spine.     These large forces were a consequence of the speed of engagement as

the weight (and number) of players involved in the scrum.                         This highlights

not   only    the     need    for    physical       preparation       of    all    forwards,    but

particularly player restraint at engagement and justifies the "crouch-

pause-engage" sequence recently introduced to 'depower' the scrum. As the

hooker is the player most at exposed to the greatest loads throughout the

scrum and subsequently most at risk he should determine the timing of

engagement of the two front-rows.

         Stability    of     the    scrum    is   an    indication     of   front-row     players'

ability to utilise their strength to transmit the force of their opponents

as well as the push of second-row and back-row players behind them in the

scrum. This appeared to be independent of size of players.                            Equally, it

reflected    risk    of    chronic    degeneration        of    the   musculo-skeletal        system

through repeated exposure to these large stresses. However, not only were

older and more experienced players better able to generate and transmit

these forces, they were also able to maintain the integrity of the scrum.

         A large proportion of individual players' efforts to generate force

was lost in their coordinated effort in a normal scrum.                            It is assumed

these forces were dissipated through players re-orientating their bodies in

the scrum situation as well as to less efficient shear forces and to the

elastic and compressive tissues in the body.                          It again reinforces the

importance of physical preparation for all forwards to better withstand the

large forces involved in scrummaging.

         Despite negative publicity surrounding the risk of serious spinal

injury in rugby union, limited research has been conducted to examine

either the mechanisms of injury or techniques implicated in causing injury.
Biomechanical   information    can    provide    systematic   bases    for   modifying

existing techniques and assessing the physical capacities necessary to

efficiently   and   safely   play    in   the   scrum.   This   will    both   improve

performance of game skills while minimising the potential for injury.

        The    value     of    the   rugby      union   scrum     has    traditionally       been     a

controversial subject.           It has been extolled by Irvine (1976) as

        "...the very heart and soul of the game and the sweat, the pain, the

aching limbs and the bursting lungs that go to make up the                               scrum     are

what earn rugby its unique place among the world's                      great games."

The scrum has also been condemned as dangerous, unnecessary, and "...the

least attractive part of the game." (Leggoe, 1984).                      To most observers and

enthusiasts the value of the scrum lies somewhere along a continuum between

these two extremes of view.

        The original purpose of the scrum was as a means of restarting play

after   an    infringement        but    this     has   been    lost    in   a    scramble     for    a

competitive edge over the opposition. Successful scrummaging has become a

powerful offensive skill in affording a base for attacking play and for

wearing   down     the       opposition.     In   addition,      scrummaging       is   used     as   a

defensive manoeuvre where the object is to deny the opposition their chance

for clean possession. A consequence of this competitiveness has been the

adoption of techniques that are contrary to the spirit of the game and in

some cases potentially hazardous to players. Specific examples of these

include "charging" together of the front-row, or conversely, front-rows

being too close together either from standing with their feet forward or

the   second-row       and    back-row     forwards     pushing    before        the   front-row      is

properly formed. Another ploy is to collapse the scrum either through a

deliberate act or from one prop bearing down on his opposite.

        The purpose of this paper is to review the available literature

pertaining to biomechanical factors which influence performance and safety

in rugby union scrummaging.              The paper will focus on the forces players are

exposed   to    during        standard    scrummaging     formations.            Implications      for

enhanced coaching techniques, the reduction of injuries, and possible law

changes will be discussed.

        Severe       spinal    injuries    resulting        from     playing        rugby,    although

comparatively few in number, have been the focus of considerable attention

in recent years. As a result it has been possible to collate details on the

incidence and mechanisms of injury and the recommendations made to reduce

or prevent such injuries.


        Statistically, a player is more likely to suffer serious spinal

injury from crossing the road or driving to or from the venue than from

playing   rugby      union     (Dedrick,   1985).      The    incidence        of    serious       spinal

injuries from playing rugby union or rugby league is low (1.48 per year for

rugby union and 1.92 per year for rugby league (Taylor & Coolican, 1987)).

These statistics were far less than the incidence of 7.6 per year for

diving accidents and 2.7 per year for equestrian (Reid & Saboe, 1991).

        The    incidence        of    cervical     spinal     injury      in    rugby        has     been

well-documented for most rugby-playing countries: Australia (Yeo, 1983;

Taylor & Coolican, 1987), Canada (Sorro, Van Peteghem & Schweigel, 1984),

England (Hoskins, 1979; Horan, 1984; Silver, 1984; Silver & Gill, 1988),

Ireland (O'Carroll, Sheehan & Gregg, 1981), New Zealand (Burry, 1979; Burry

&   Gowland,     1981),       South   Africa     (Scher,     1977,    1987),        USA    (Micheli    &

Riseborough, 1974) and Wales (Williams & McKibbin, 1978, 1987).                              From these

data,   more     than     50    reported   injuries        could     be   directly         related    to

scrummaging, and of this number, only one was identified as occurring to a

player other than in the front-row.                Furthermore, injuries were identified

as occurring equally between the props                 and the hooker.              The data did not

distinguish between loose-head or tight-head prop.

        Age    and    experience      appeared    to   be    related      to    the       incidence   of

cervical spinal injury among front-row players.                      Burry (1979) reported the

six props injured in scrums in New Zealand between 1973 and 1978 were under

20 years of age.          Williams & McKibbin (1978) reported only one of five

front-row players injured was older than 20 years.                        Similarly, Yeo (1985)
reported only three of eight front-row players injured were older than 20


         In an effort to reduce the risk of cervical spinal injury from

scrummaging, a variety of law modifications have been proposed.                       One

modification suggested by several administrators has been a return to

earlier forms of scrummaging.             When the subject of risk of injury is

discussed    one   frequently     hears    administrators    and   former   rugby    union

players declaring 'it didn't happen in my day'. The claimed lack of injury

in earlier times may possibly be due to the manner in which the game was

played in that era, or more than likely the absence of public awareness of

the risks involved or lack of media coverage. To refute this perception of

the game being safer in the past, O'Connell (1954) reported six cases of

fracture-dislocation     of     cervical    vertebrae   amongst     rugby   players    in

Leinster (Ireland) in the period 1934-1954. In three cases death occurred

within 24 hours.        An even earlier reference drawing attention to the

violence among rugby players was the report of 71 deaths in Yorkshire

during the three seasons between 1890 and 1893 (The Wakefield Express,


         A very high proportion of all rugby injuries result either from poor

technique or from illegal or foul play (Wessels, 1980).              It has been shown

that the majority of incidents resulting in spinal injury in American

football were a result of a deliberate action ("spearing" or head-on

tackling) which placed the cervical spine at risk (Torg, Truex & Marshall,

1979). Scher (1981) alluded to this risk with respect to direct vertex

impact injuries during tackling and scrummaging in rugby union. Law changes

were introduced in all Australian matches for the 1988 season with the

intention of increasing player safety in several aspects of the game

(Australian    Rugby   Football    Union,    1992).     In   conjunction    with    strict

enforcement of the modified scrummaging laws there have been no instances

of scrum-related serious cervical spinal injury in Australian rugby union.

         The effectiveness of these law changes still depend on the referee's

discretion in situations, such as, what constitutes charging at engagement,
when to terminate the scrum if it collapses, or when a player's head is

"popped" out of the scrum.           A referee may not be aware of a player who is

tightly bound in a scrum and subsequently at risk of sustaining serious

cervical spinal injury. By allowing play to continue the referee may be

subjecting the player at risk to unnecessary trauma.                A call for referees

not to prolong the scrum has been made by several authors. Scher (1987) and

Cohen & Siff (1979) recommended a time limit be imposed on the duration of

scrums as a safety measure. They suggested prolonged scrummaging would

increase the risk of a scrum collapse and            expose players to flexion injury

of the cervical spine.


         The tightly bound rugby scrum would appear to be a mechanism that

places the cervical spine at risk of injury (Scher, 1977).                     Injury can

occur during formation of the scrum when front-rows "charge in" and a

player does not have his head correctly aligned. Typically this action

could result in hyperextension of the neck as the head is "popped" out of

the scrum, or more commonly, from compression and flexion/rotation of the

cervical spine when a scrum collapses and the scrum or opponents continue

to   push.       Other    potential     injury    situations     include    "wheeling"    or

"screwing" a scrum, and the situation when players rotate their necks to

look at the incoming ball.           These techniques can result in a combination of

lateral flexion and rotation of the neck causing excessive horizontal


         Scher   (1977,      1982)    observed   that   all     players    injured    during

scrummaging      sustained    similar    injuries.      These   injuries    were     fracture

dislocations with bilateral locking of facets at levels of C4/5 or C5/6.

Injuries of this nature are considered the result of flexion/ rotation

violence which leads to rupture of the posterior spinal ligaments and

vertebral dislocation (Scher, 1977).             The aetiology of other scrum related
cervical spinal injuries have indicated that hyperextension of the neck can

result from a player being lifted out of a scrum ("popped") and direct

compression     can    occur    from    a    clash   of   heads    with   the    opposite      prop

(Williams & McKibbin, 1978).

         Use of protective gear such as helmets can decrease the effects of

impact loading by spreading the load and increasing the duration of impact.

However it is considered almost impossible to protect against the effects

of acceleration/deceleration loading of the spine by contiguous structures.

Immobilisation of the cervical spine by use of a cervical collar has been

proposed by Naylor & Neal (1988) as a means of preventing cervical spine

injury in rugby. All studies evaluating various methods of neck fixation

including hard and soft cervical collars have demonstrated poor fixation,

limited     cervical        motion    and     questionable       shock-absorbing        properties

(Watkins, 1986). Only when the base of the collar was extended and fixed

firmly to the chest, as in a cervicothoracic orthosis, could cervical

motion be restricted. Fixing the head to the chest however is contrary to

the demands of rugby players who require full range of motion of the

cervical spine.        For all practical purposes, cervical spine injuries can

therefore be considered a function of technique and cannot be measurably

prevented by equipment (Burstein, Otis & Torg, 1982).


         The scrummaging technique considered to be most effective has not

been well defined and is based almost exclusively on anecdotal evidence. A

search    of    popular      literature       revealed    the     availability     of    numerous

"how-to-do-it" articles and book chapters written by current coaches and

former exponents of the game of rugby union. Several studies referred

indirectly to the technique of scrummaging and were primarily directed

towards the injury potential of rugby scrums (Cohen & Siff, 1979; Hodge,

1980;    Sumner,    1980;     Stubbs,       1981;   Rodano   &   Pedotti,   1988).       All   data

presented      in   these    papers    were    limited    to     the   forward   force    without

reference to vertical or lateral shear forces experienced by front-row
players.     Only   recently    have    the    three   orthogonal   components        of   force

experienced during scrummaging been investigated (Milburn, 1987, 1990a,b).

Vertical and lateral shear forces have been recognised as an important

factor in causing the traumatic injury associated with scrum collapse along

with   the   chronic     degeneration    of    discs,   vertebral       bodies,     ligamentous

structures, and apophyseal joints.             The repeated stress the cervical spine

is subjected to particularly among front-row players increases the risk of

hyperextension injury to the cervical spinal cord (Scher, 1983).

         It is the obvious lack of information, and the risk of catastrophic

spinal    injury    as   a   consequence      of   incorrect   technique      and    inadequate

knowledge which prompted extensive investigations into the biomechanics of

various scrummaging techniques (Milburn, 1987, 1989, 1990a,b).


         Front-row players are most vulnerable to serious cervical spine

injury during engagement of the scrum when front-rows "charge in", when the

scrum collapses, or when a head is "popped" out of a scrum as the two

front-rows    come   together.         These   events   are    caused    by   any    one   or   a

combination of the lateral, vertical or forward forces acting individually

or collectively on front-row players. Once the scrum is settled players

exert a second sustained shove in an attempt to retain possession or

disrupt the opposition's possession.                An understanding of the nature of

forces likely to produce these injury situations during engagement and

throughout the scrum would provide a basis for eliminating dangerous or

counter-productive aspects of the game.

         All scrums examined by Milburn (1987, 1990a,b) using an instrumented

scrum machine showed a large forward impulsive force on engagement produced

as a result of stopping forward motion of the scrum.                     Typically this was

followed by a considerable drop in force on all players within the first

second of scrum.         This drop in transmitted force was probably due to the

observed "compression" and realignment of the scrum following engagement.

Within the first second of the scrum being formed the motion of players and
forces exerted became stabilised.         Mean data for the three components of

force at engagement and during sustained scrummaging for teams of differing

ability levels using a standard scrum formation are presented in Table 1

(Milburn, 1990a,b).

Table 1.   Mean Force Distribution Across the Front Row at Engagement and

During Sustained Scrummaging (Newtons)

                  Engagement                        Sustained

           Loose Hooker   Tight   Total     Loose Hooker    Tight    Total
           Head           Head              Head            Head

Lateral Force

H.S        320    -1410   1240     150        640    1520    780     3040

Univ.       90      70     570     730         80    590     840     1510

Club       363    1350     700    2413        255    2150    688     3093

Under 23   -556     30      68    -551        -44    -195       18   -232

Internat   -22      55      52      85        155    107        77   340

Vertical Force

H.S.       -810    380    -510    -940        630    -130   -310     190

Univ.      -490   -600     930    -160       -380    330     660     610

Club       -280   -413    -175    -868       -113      0     -38     -151

Under 23   -139   -296    -262    -959        355    709     352     1370

Internat   498    1307     463    2268        309    581     415     1305

Forward Force

H.S.       1500   2120    810     4430       1090    1650    630     3370

Univ.      1070   3380    2090    6540       1150    2070   1390     4610

Club       1345   3160    1125    5630        750    2300   1250     4300

Under 23   1723   2837    1580    6140       1809    1878   1221     4908

Internat   2484   3778    1710    7982       1494    2751   1516     5761
         Lateral force, apart from being inefficient in terms of the aims of

contemporary scrummaging, has been identified by Scher (1983) as causing

premature degeneration in the cervical spine. Any lateral shear forces at

engagement      would   therefore       have    implications      for   player      safety      as   to

whether they were considered excessive or likely to place players at risk.

Similarly, lateral shear force during the sustained or second shove could

indicate flaws in technique or potential limitations of particular scrum

formations or binding combinations.

         These forces were relatively small compared to the magnitude of the

overall forces acting on these players and should not normally pose a

problem at engagement (Milburn, 1990a). However, players who were not set,

ready for engagement, fatigued, injured, or propping against a player who

was intent on "boring-in" on his opponent could be at risk of clashing

heads with his opponent and exposing himself to vertex impact injury.

         Taylor & Coolican (1987) reported seven serious spinal injuries were

a result of the scrum collapsing and a further four from pushing after the

scrum    had    already   collapsed.           Scher   (1991)     reported    all    16    injuries

sustained in scrummaging occurred due to collapse of the scrum.                              The New

Zealand Rugby Football Union (1979) identified that front-rows who tended

to stand too close to each other and second-row and back-row forwards who

applied the push before the front-row was properly formed contributed to

the risk of scrum collapse at engagement. The alignment of players and

therefore       the   tendency    for    a     scrum   collapse    at     engagement      would      be

represented by the vertical direction of forces acting at engagement.

         All of the front-row forwards would be expected to experience a

downward force at engagement due to a proportion of their body weight being

forward of their base of support when they moved forward and off-balance to

engage the opposition.           The magnitude and direction of this force would be

modified by the strength and direction of the leg drive. Milburn (1990a,b)

recorded    a    downward   force       acting    on   front-row     forwards       in    all   scrum

combinations at engagement. Although these forces were not large in their

own     right    (approximately     half       body    weight),    they    would     assume      some

significance in contributing to the risk of scrum collapse at engagement.
If players were not properly aligned, mis-matched in terms of strength or

experience, fatigued, charged each other at engagement, or were being

pulled downwards by their opposing prop, these downward forces would be


         Even if care was taken in engaging the two scrums, forward forces

experienced would still be substantial.           A mean recording of 5922 N for

total forward force at engagement was recorded for all scrums tested which

represented an average force of more than half a tonne shared across the

front-row.     Values ranged from 3470 N for a novel scrum formation to 7980 N

recorded by an international team.          The force was an impulsive force of

very   short   duration   necessary   to   stop   the   motion   of   the   scrum.   An

impulsive force would be larger in stopping a scrum with a greater mass

(weight), as would be the case with the international team compared to the

Club team.     An impulsive force would also be larger if there was a greater

change in the motion (speed) of the scrum over a short time interval, as

would occur on impacting a rigid scrum machine. Therefore, the engagement

force was as much indicative of the speed at which the scrum engaged the

scrum machine as it was of the ability of a scrum to apply forward force

through coordinated muscular action.

         Data obtained were less than the peak forward force reported by

Hodge (1980) of 8988 N for a low scrum (front-row packing as low as

possible) using a university forward pack.          These data are consistent with

the hypothetical maximum of 7523.5 N reported by Sumner (1980) determined

by geometrically summing the forces exerted by a single front-row player in

a fully extended pushing position.          The combined force of two opposing

forward packs on engagement would therefore be close to the "one and a half

tonnes" (14955 N) of force estimated by Stubbs (1981) that was often

reported as the load placed on front-row players (Burry, 1979; Scher,


         These forces are substantial and represent a considerable risk of

cervical spinal injury to an ill-prepared player.                The magnitude of the

impulsive force experienced by individual players at engagement was found

to exceed the threshold for injury to the spine in flexion (2000 N: Mertz &
Patrick, 1971), as might happen in a scrum collapse.             This force was also

at the lower end of the threshold for injury in direct compressive loading

(between 3338 and 4451 N; Burstein et al., 1982), as might happen in a

clash of heads. This would indicate the importance of correct alignment of

the head, neck and trunk, along with adequate strength to maintain body

position during engagement.


       The sustained or second shove is considered to be a measure of a

team's effectiveness in scrummaging. It represents their ability to ensure

clean possession or to disrupt their opponent's possession. Similarly it

represents a scrum's capacity to withstand the efforts of their opponents

in "holding" their position against all efforts.

       Hodge (1980) compared the forward force produced in eight different

binding or scrum formation techniques.             Results showed the full scrum

exerted greatest force (8988 N) in the low scrum formation (front-row

packing as low as possible).           Other techniques including the double push

and staggered scrum techniques produced progressively less forward force.

Sumner (1980) examined the forward forces generated by players in pushing

against a floor mounted force plate while using different body alignments.

The ideal pushing position was found to be with the head, trunk and legs in

alignment and generally the greater the angle between the trunk and thigh,

the greater the forward force. The resultant force vector through the feet

was   forwards   and   upwards   and    never   greater   than   27   degrees   to   the


       Of the remaining studies evaluating scrummaging technique, Cohen &

Siff (1979) measured the forward force exerted by the front-row only using

an instrumented scrum machine.         The maximum static force exerted was 2800 N

and from this a conservative estimate of the static force exerted by the

whole scrum was 7000 N. In comparison, Stubbs (1981) reported the combined

forward force exerted by the whole scrum totalled 900 lbs (4000 N).                  The

combined front-row and second-row sub-unit provided 650 lbs (2893 N), with
400 lbs (1780 N) provided by the front- row only. Stubbs estimated the

total combined force of an international forward pack would be 1500 lbs

(6676 N).       The much larger forces experienced at engagement were due to

stopping the motion of opposing forward packs.              In comparison, the static

nature     of   the   sustained   second    shove   would   be    expected     to   produce

considerably less force than the previously reported "one and a half

tonnes" of force at engagement.

         All force data previously reported (Cohen & Siff, 1979; Hodge, 1980;

Sumner, 1980; Stubbs, 1981) have been limited to the forward direction

without reference to the vertical and lateral shear forces experienced by

the front-row players.       One other paper (Rodano & Pedotti, 1988) utilised

two floor-mounted force plates to determine the three components of force

applied through the legs. Two variations of leg position were tested during

sustained and "impulsive" shoves. It was assumed the total horizontal

thrust recorded        (left leg plus right leg) was equal to the horizontal

(forward) force at the shoulder.           Vertical force was not included in their

analysis but was recognised both as a tactical ploy to destabilise the

scrum and as an indicator of reduced efficiency.            The lateral component was

not considered as the direction of the leg thrusts were opposite and would

tend to "stabilise around zero" (p. 477).

         At the individual level, Rodano & Pedotti (1988) found the impulsive

shove produced higher peak forces and also required greater individual and

team coordination skills than a continuous shove.               There was no difference

in   the   player's    "steady-state"      leg   thrust   (p.    482)   when   either   the

impulsive (mean of 1529 N) or continuous shove (mean of 1520 N) were used.

It was noted no significant correlation existed between a player's body

weight and leg thrust or between joint angles in the leg and magnitude of

thrust. No attempt was made to geometrically sum the individual players'

forces to produce an estimate of total scrum force.

         A small lateral force was indicative of the stability of a scrum

during the sustained or second shove whereas large lateral forces were

indicative of inefficient scrummaging technique.                 In all scrums tested,
Milburn (1990b) reported a force across the front-row directed towards the

loose-head prop and the incoming ball. In the traditional scrum formation

the loose-head prop typically experienced a greater proportion of forward

force   transmitted    (1551   N   compared    to   1288    N   in    the   standard   scrum

formation). This result was consistent with Hodge (1980), who reported the

loose-head prop transmitted 60 percent of the forward force.                      However,

during engagement Hodge reported the tight-head carried proportionately

more forward force.      This load shift is indicative of the lateral shear

experienced by front-row players.

        In the majority of scrum combinations examined by Milburn (1990b),

the vertical force during the sustained or second shove was directed

upwards. When the hooker bound over (and forward of) the shoulders of the

props, he was potentially vulnerable to being "popped" upwards as the

second-row applied forward force. Use of the crotch-grip by the second-row

did not adversely affect the vertical force acting on the front-row nor

increase the risk to players.


        A comparison of the forces experienced by front-row players using

different binding techniques was undertaken by Milburn (1987).                     The hip

binding (second-row binding around the hip of the prop and grasping his

jersey above his hip) and crotch binding (second-row binds onto the prop's

jersey by reaching between his legs) techniques were compared. Milburn

found   the   crotch   binding     technique   added       to   the   downward   force   at

engagement and was considered to increase the risk of scrum collapse at


        The additional downward pull on the prop's jersey with the crotch

grip was assumed responsible for an increased downward force recorded on

the props.     In comparison, no differences due to binding technique were

found in vertical force data during sustained scrummaging (Milburn, 1990b).

Although not able to record vertical or shear forces, Hodge (1980) found
there was no difference in the forward force exerted during sustained

scrummaging by using the crotch binding technique compared to hip binding.

         Milburn (1990b) examined the effect of adopting the "over" method of

binding by the hooker (binding over the shoulders of the props) which

resulted in his shoulders being forward of the props' shoulders.                        When in

contact with the scrum machine, the hooker's shoulders tended to carry a

proportion of the props' load which was reflected in a larger forward force

on the hooker. Conversely, with an "under" binding (binding under the

shoulders of the props) this force would be partially spread across the

props' shoulders and be reflected in a lower forward force recorded by the


         Examination of different binding methods without consideration of

grip combinations indicated scrums in which the hooker bound over the props

produced more forward force (4577 N) than when the hooker bound under the

props.      When   the    grip   used     by    the      second-row    was   examined    without

consideration of hooker binding combinations, the crotch grip produced more

forward force (4614 N) than did the hip grip (4220 N).                       These results are

consistent with greater force production in more familiar combinations.


         Very little information was available on the contribution of various

sub-units of the scrum to the total forward force produced.                       Hodge (1980)

reported that reducing scrum numbers to six forwards by eliminating the

flankers produced 28 percent less forward force.                  Eliminating the lock did

not further reduce the forward force.                    A similar result was obtained by

Milburn    (1990a)       where   data     from       a    comparison    of    scrum     sub-unit

contributions showed flankers contributed 20 - 27 percent to the forward

force and the lock (Number 8) contributed little additional forward force.

The contribution of sub-units of the rugby league scrum has also been

determined    by     Milburn     (1989)    in    a       comparison    of    contemporary   and

recommended binding techniques.           Results from the rugby league scrum showed

the second-row forwards' role appeared to be one of supplying forward
force. The role of the back-row forward (lock) in a rugby league scrum was

not to provide forward force but to reduce shear forces and stabilise the


         Cohen & Siff (1979) used an instrumented scrum machine to measure

forward force exerted by the front-row only.                   The maximum static force

exerted was 2800 N which represented 40 percent of their predicted maximum

force for the whole scrum. Stubbs (1981) reported a contribution of 1780 N

for   the   front-row   only   and    2893    N   for   the    front-row   and   second-row

sub-unit. The front-row contribution corresponded to 45 percent of the 4000

N reported for the whole scrum. From these data it could be deduced the

second-row contributed 1113 N or 28 percent of the total forward force as

compared to 46 percent determined by Milburn (1990a).

Table 2. Sub-unit Contributions to Forward Force Averaged from Previously

Reported Data

                               Mean          Percent            Range

         Front-row only        2623 N             42          1780 - 3290 N

         Second-row            1617 N             37          1113 - 2120 N

         Flankers              1339 N             25          720 - 2517 N

(Source: Cohen & Siff, 1979; Hodge, 1980; Milburn, 1990a; Stubbs, 1981;

Rodano & Pedotti, 1988)

         It is assumed these force data were for a sustained, continuous

shove and did not reflect forces at engagement. Previous investigations

have shown the forward forces at engagement were substantially greater than

those recorded for the sustained shove. In one instance, the greatest force

recorded at engagement was the front-row of the lightest forward pack

(Milburn, 1990a).       In this trial the front-row engaged the scrum machine

with the highest velocity which further reinforced the need to eliminate

the charging together of scrums at engagement as well as players charging

blindly into rucks and mauls.
         Rodano & Pedotti (1988) isolated continuous forward thrust produced

by each forward.      It was found that each prop was capable of producing a

forward force of between 1520 N and 1844 N. The second-row players were

each capable of producing forces of 1137 N - 1520 N, flankers of between

1667 N - 1746 N, and the lock a force of 1844 N.             Even if the minimum value

was    taken   for   each    player,   the    total    forward    force   (excluding   the

contribution of the hooker) amounted to 10492 N.             A considerable proportion

of the force produced by forwards other than those on the front-row would

be     dissipated    in     the   elastic     and   compressive     components   of    the

musculoskeletal systems of players forward of them in the scrum. Additional

amounts of the resultant forward force would be "lost" as lateral and

vertical components of the force due to body orientation or the way players

packed into the scrum.

         Milburn (1990b) estimated the contribution of various sub-units of

an International forward pack by subtracting the total forward force on all

three front-row players from the total for a complete scrum (Table 3). The

contribution to the total forward force output of the scrum for each

sub-unit was: front-row (38 percent), second-row (42 percent), flankers (8

percent), and lock (12 percent).             The data on the three back-row forwards

(flankers and lock) differed from that previously reported when they were

considered separately (Milburn, 1990a), but their collective contribution

(20 percent) was consistent with other research.

Table 3. Sub-Unit Contributions

                                       Engagement      Sustained

      Front-row only (N=3)                      6569             2168

      Front-row plus second-row (N=5)           6408             4610

      Full scrum minus flankers (N=6)           ----             5329

      Full Scrum (N=8)                          7982             5761
      It is worth noting the extent to which data on individual player

contributions reported by Rodano & Pedotti (1988) differed from their

effect on the forces transmitted by the front-row. Only their data on the

second-row (2657 N) was consistent with that observed by Milburn (1990b).

The collective thrust of the three back-row forwards (5257 N or 50 percent

of the total scrum) substantially over-estimated the contribution recorded

for the back-row forwards in Milburn's study (1222 N or 20 percent).

      A comparison was also made of the combined maximum forward force

data recorded for individual members of the scrum and that obtained for the

complete scrum.    The sum of the individual player's maximum forward force

was 17725 N (approximately one and three-quarter tonnes) which represented

more than three times that recorded for the full scrum (5761 N).                Not

included in the calculation of this large force was the contribution of the

hooker. These results reflect the extent to which individual player's

efforts are attenuated by the "give" in the body's structures as well as

some forward force being "lost" to less efficient lateral and vertical

components of force. Similarly, the shape and orientation of binding and

pushing surfaces differed between the scrum machine and the intact scrum

which may have hindered a player's ability to apply force.             However, it

again reinforced the need for physical preparation of all forwards to be

able to withstand the large forces involved in scrummaging.


      A comparison was made between the differences in standard of the

players and the forces experienced at engagement and during the sustained

second shove (Milburn, 1990a,b). Mean data for High School, University,

Club, Under 23, and International teams using their usual binding technique

are presented in Table 1 along with the proportion of the forward force

carried by each player.

      Experience   has    been   considered   a   major   attribute   of   front-row

players.   This was reflected in the increased lateral stability recorded

for higher level front-row players which in turn illustrated their ability
to maintain the integrity of the scrum. Lateral stability is also an

indicator of a team's ability to transmit forward force more efficiently.

Conversely, vertical forces acting on the front-row at engagement were not

found to be related to the experience level of players.                    A net negative

(downward) force at engagement was recorded on all but the International

front-row.     With High School, Club and Under 23 teams, this force was

substantial. However, a trend towards greater positive (upward) vertical

forces was recorded for higher-level teams during the sustained second

shove.       The   risk   of    scrum    collapse     would    therefore   be     less    with

higher-level teams.       Any collapse that occurred would likely be a result of

an accident or technical fault (for example, a player slipping and losing

his footing), or through a deliberate act.

         As would be expected, in improvement in a team's ability to apply

forward    force    corresponded      with     an   improvement   in   performance       level

(Milburn, 1990b). In some instances, there were negative (downward) forces

acting on individual players during the sustained second shove indicating

the risk of a scrum collapse would be greater for lower-level teams.                     These

data     further   reinforce    the     need   to   maintain    correct    body    alignment

(shoulders above the level of the hips) throughout the duration of the



         Players are often called upon to play in a position different to

their more familiar role. This be caused by an injury, players being

unavailable, or being sent off during a game. It is also a requirement of

the Under 19 Law Variations (Law 20 (5) Note ii) that "Should a team be

unable to provide competent players for the front or second row, the

referee should not permit the game to proceed" (Australian Rugby Football

Union, 1992, p. 189).          Therefore, competent players must be available to

accommodate changes in player personnel.              However, there is also a need to

assess the effect on scrum stability when the composition of the scrum is

changed.     Torg (1991) suggested injury was more likely to occur if forward
packs were not familiar with each other and "...uncertain of their combined

course of action in the event of a collapsed scrum" (p. 194).

       Milburn (1990b) observed several consequences of changing players'

positions were observed in the forces recorded on front-row players.                           Most

notable was the normally slight (-165 N) left-directed force acting across

the front-row increased substantially (-1471 N) at engagement when the

props were reversed. The force acting on the hooker also changed direction

from   right-directed    to    a   left-directed       force.    These   changes        in    force

distribution would have a consequential unsettling effect on the front-row

and increase the risk of players being out of alignment at engagement in

the scrums immediately following a change in personnel. There was no

substantial change in magnitude of the lateral force acting across the

front-row    during   the     sustained     second     shove    with   changes     in    playing

positions.    However, reversing the props had the effect of reversing the

direction of this force which would further unsettle the scrum.                     The effect

of this change in direction would be compounded if the opposing scrum was

using its normal player configuration.

       The consistent trend of downward forces recorded during engagement

was also apparent for all changes in player personnel.                    Similarly, during

the sustained second shove, all front-row combinations tested were able to

maintain an upward force despite changing position or personnel.                        However,

changes in second-row and back-row player personnel created differences in

vertical    force   direction      acting   on   the    front-row.       In   one    instance,

changing the second-row caused a downward force to be acting on the front-

row. The implications are that changes in personnel other than in the

front-row also appeared to have the effect of destabilising the scrum and

increasing the risk of collapse during the second shove.                         Care should

therefore    be   exercised    during     scrummaging     following      changes    in       player


       The forward force at engagement differed very little from normal

scrum engagement when front-row and second-row personnel were varied.                         This

becomes important when considered in conjunction with the large differences

in lateral force if front-row players are reversed.                    The result of these
changes further emphasises the need for caution at engagement when forwards

are playing in unfamiliar positions.                Similarly, changes in the front-row

did not adversely affect sustained forward force.                   This further indicates

the contribution of front-row forwards in generating forward force as well

as their ability to transmit the force of other players in the scrum.

However, changes in the second-row and flank forwards had the greatest

influence on forward force during the second shove. In both situations the

total forward force was approximately 1000 N less than for a normal scrum.


        Despite the large number of players involved in the game of rugby

union and the negative publicity given to the incidence of serious injury,

there is a paucity of research information related to the manner in which

the    game   of   rugby   union     is   played.       This   is   surprising   given    the

recognition afforded International success and the effectiveness of Law

changes instigated by the Australian Rugby Union based on support provided

by previous research (Milburn, 1990b).               The ability to accurately describe

the mechanism responsible for a particular injury is a necessary precursor

to    preventing   that    injury.        Equally    important,     research   can   discount

erroneous concepts, and if unchallenged, can become established as fact

(Otis, Burstein & Torg, 1991).               Research can provide a systematic and

objective understanding and reasoning behind a recommendation or adoption

of a particular technique, and why one technique is more favourable than


        The current programme of research instigated by the author (Milburn,

1987, 1990a,b), and the epidemiological evidence of injury in rugby union

(Burry & Gowland, 1981; Scher, 1987; Silver, 1984;                     Taylor & Coolican,

1987) have all indicated serious spinal injury in rugby union is a function

of technique. Similarly, research programmes in the United States found

injuries in American football were a consequence of incorrect or unsafe

technique.     Development and implementation of rule changes has successfully

reduced the incidence of serious spinal injury.                     Therefore, information
presented from biomechanical analyses can provide administrators, coaches

and players with the necessary theoretical bases for player selection,

coaching, law modification and ultimately player safety.


Australian Rugby Football Union. The Laws of the Game of Rugby                   Football,

Sydney: Australian Rugby Football Union, 1992

Burry,     HC. Report to the Medical Advisory Committee to the New Zealand

         Rugby Football Union (Inc.) Council on Injuries to the Cervical

         Spine, 11, 1979

Burry, HC, Gowland, H. Cervical spine injury in rugby football: a New

         Zealand survey.         British Journal of Sports Medicine, 15, 15-19,


Burstein,       AH,   Otis,   JC,    Torg,   JS.    Mechanisms   and    pathomechanics   of

         athletic     injuries      to the cervical spine. In Torg (Ed)            Athletic

         Injuries to the Head, Neck and Face. Philadelphia: Lea & Febiger:

         139-154, 1982

Cohen,    I,    Siff,   M.    Increased safety in rugby scrums.               South African

         Medical Journal, 56, 625, 1979

Dedrick, JD. How to avoid spinal injuries playing sport. Sports                  Monthly,

May, 2-8, 1985

Hodge, KP. Spinal Injuries in Rugby Scrums. Research Papers in Physical

         Education, Vol. 2, No. 4, University of Otago, 82, 1980

Horan, FT. Injuries to the cervical spine in schoolboys playing rugby

         football. Journal of Bone and Joint Surgery, 66-B, 470-471, 1984

Hoskins, T. Rugby injuries to the cervical spine in English schoolboys.

         Practitioner, 233, 365-366, 1979

Irvine, D. The Joy or Rugby, Wellington: AH & AW Reed.

Leggoe, J. Letter to the Editor, The Australian, 6 June, 1984

Mertz, HJ, Patrick, LM. SAE Paper 710855, Proceedings 15th Stapp Car

         Crash Conference,          207 (in Goldsmith, W. Some Aspects of Head and

         Neck     Injury      and     Protection,     in   Akkas       (Ed)   Progress   in
        Biomechanics Alphen aan den Rijn, The Netherlands: Sijthoff and

        Noordhoff, 1979. pp. 333-378), 1971

Micheli,      LJ,    Riseborough,           EM.    The       incidence       of     injuries        in    rugby

        football Journal of Sports Medicine and Physical Fitness, 2, 93,


Milburn, PD. A comparison of the mechanics of hip and crotch binding

        techniques          in    rugby     union       scrummaging.         Australian           Journal     of

        Science and Medicine in Sports, 19, 3-9, 1987

Milburn,      PD.    Mechanics         of   a     Rugby      League      Scrum.        In    Morrison       (Ed)

        Proceedings          VIIth      International          Symposium          of     the      Society    of

        Biomechanics in Sports, 61-72, 1989

Milburn, PD. The kinetics of rugby union scrummaging. Journal of Sports

        Sciences, 8, 47-60, 1990a

Milburn, PD. The Biomechanics of Rugby Scrummaging.                                    Unpublished report

        to Australian and New Zealand Rugby Football Unions, 1990b

Naylor,    DJ,      Neal,    R.    Abstract:       Analysis         of   a   neck      brace      for    rugby.

        Australian Journal of Science Medicine in Sport, 20, 24, 1988

New Zealand Rugby Football Union. Report of the Medical Advisory Committee

on Injuries to the Cervical Spine, 1979

New Zealand Rugby Football Union.                       Injury Prevention in Rugby Football:

        Situation Paper No. 1 (10th May), 1984

O'Carroll, PF, Sheehan, JM, Gregg, TM. Cervical spine injuries in rugby

        football. Irish Medical Journal, 74, 377-379, 1981

O'Connell, TCJ. Rugby football injuries and their prevention. A review                                      of

600 cases. Journal of Irish Medical Association, 34(199), 20-                                      26, 1954

Otis,   JC,    Burstein,         AH,    Torg,     JS.        Mechanisms       and      pathomechanics        of

        Athletic      Injuries         to the Cervical Spine. In Torg (Ed)                           Athletic

        Injuries to the Head, Neck and Face. (2nd ed.). St. Louis: Mosby

        Year Book, 1991

Reid, DC, Saboe, LA. Spine injuries resulting from winter sports.                                            In

        Torg     (Ed)   Athletic          Injuries      to    the    Head,        Neck      and   Face.     (2nd

        ed.). St. Louis: Mosby Year Book, 1991
Rodano,    R,    Pedotti,   A.    A   quantitative        analysis      of   thrust    and       motor

         coordination in a rugby scrum.            In Reilly et al. (Eds) Science                 and

Football. London: E & FN Spon, 476-486, 1988

Scher, AT. Rugby injuries to the cervical spinal cord.                             South African

         Medical Journal, 51, 473-475, 1977

Scher,    AT.    Vertex   impact      and    cervical     dislocation        in   rugby    players.

         South African Medical Journal, 59, 227-228, 1981

Scher, AT. "Crashing" the rugby scrum - an avoidable cause of cervical

         spinal injury. Case reports.              South African Medical Journal, 61,

         912-920, 1982

Scher,     AT.     Serious cervical spinal injury in the older rugby player.

         South African Medical Journal, 64, 138-140, 1983

Scher, AT. Rugby injuries of the spine and spinal cord.                            In Torg (Ed)

         Clinics in Sports Medicine, 6(1), 87-99, 1987

Scher,    AT.    Catastrophic     rugby      injuries     of   the     spinal     cord:    changing

         patterns of injury. British Journal of Sports Medicine, 25(1),                           57-

60, 1991.

Silver, JR. Injuries to the spine sustained by rugby. British Medical

         Journal, 288, 37-43, 1984

Silver,    JR,    Gill,   S.    Injuries      of   the    spine   sustained       during     rugby.

         Sports Medicine, 5, 328-334, 1988

Sorro, OM, Van Peteghem, PK, Schweigel, JF. Cervical spine injuries on

         rugby   players.      Canadian      Medical     Association     Journal,      130,       735-

         736, 1984

Stubbs, DA. The Sunday Times, 10 May, 1981

Sumner, N.       An Investigation into Forces and Pressures in a Rugby Scrum

         and Particular Injuries that Occur in the Spine. Unpublished BSc

         Report, Liverpool Polytechnic, 1980.

Taylor,     TFK,      Coolican,       MRJ.     Spinal-cord        injuries        in   Australian

         Footballers,     1960-1985.        Medical    Journal    of    Australia,        147,    112-

         118, 1987.
The   Wakefield       Express.      8th     April, 1893.       cited in Dunning & Sheard.

        Barbarian,         Gentlemen      and     Players.    Canberra:   Australian      National

        University Press, 1979

Torg,   JS.         Editor's    Note (in Taylor, TFK, Coolican, MRJ Spinal Cord

        Injuries in Football - Rugby Union, Rugby League, and Australian

        Rules.        in Torg, (Ed) Athletic Injuries to the Head, Neck and

        Face (2nd ed) St. Louis: Mosby Year Book, 1991

Torg, JS, Truex, RC,                Marshall, J. Spinal injury at the level of the

        third       and    fourth     cervical     vertebrae    from   football.    Journal       of

        Bone and Joint Surgery, 59A, 1015-1019, 1977

Watkins,      RG.     Neck     injuries      in    football    players.   Clinics    in     Sports

        Medicine, 5(2), 215-246, 1986

Wessels, LGD. Rugby injuries in South Africa. South African                          Sports

Medicine Journal, 8, 14-18, 1980

Williams,     JPR,        McKibbin,    B.    Cervical    spinal    injuries   in    rugby   union

        football. British Medical Journal, December, 23-30, 1978

Williams, P, McKibbin, B. Unstable cervical spine injuries in rugby -                         a

20-year review. Injury, 18, 329-332, 1987

Yeo, JD. The Mechanisms of Neck Injury. Paper presented at the Solution

        to Neck Injuries in Football public forum, Sydney, 1983

Yeo, JD. Acute Royal North Shore Injury Admissions to Spinal Injuries

        Unit from 1980-1984. Unpublished report, 1986