Among the questions an archer asks early on is one that sounds elementary: which muscles work when I shoot, and how hard. The answers you hear on the range are images rather than measurements. You shoot with your back; the drawing arm is a hook and the hand does nothing; at the release the back keeps pulling. These images are useful for teaching, but muscles can also be measured, and the studies on archers that have done so give a fairly precise picture on some points and a very uncertain one on others.
On which muscles switch on, the studies largely agree: the muscles between the shoulder and the shoulder blade on the drawing side work hard, the bow arm is not passive, and comparisons between levels suggest that, as the level rises, the work moves away from the forearm and the arm towards the shoulder blades. On how hard they work, the figures for the same muscle can vary four- or fivefold from one study to the next, and to see why it helps to start with how they are measured.
How muscle work is measured
Surface electromyography records the electrical signal the nervous system sends to a muscle. The raw signal, in microvolts, says little on its own, because it depends on the skin, the position of the electrode and the size of the muscle. So almost every study expresses it relative to a reference test, usually the maximal voluntary contraction, in which the archer pushes as hard as possible against a fixed resistance: when a study says a muscle is working at 40%, it means 40% of the signal recorded in that test.
This brings two weaknesses worth keeping in mind for every figure in this article. The first is that the maximum depends on how it is obtained, meaning the position of the arm, the resistance and how hard the person tries, so the same muscle can come out far more or far less active depending on the test. The second is the time window: the average over the whole shot, the average over the second before the clicker and the average over the hundred milliseconds around the click are different quantities, because activity rises around the clicker. When two studies report figures far apart for the same muscle, the gap often lies here.
The drawing arm while aiming
The highest figures come from a Turkish experiment on four elite recurve archers, 108 arrows at 18 m, with electromyography of the drawing arm normalised to the maximal contraction. In the second before the clicker fell, the posterior deltoid was at 85.4%, the finger flexors at 83.2%, the finger extensors at 82.5% and the trapezius at 72.4%. If these percentages hold, the drawing arm is not a passive hook while aiming: it works close to its maximum, fingers included.
The same data were published twice and count as a single experiment. The claim that while aiming, the drawing arm's muscles work at over 80 per cent of their maximum, and at the release almost all of them switch off except the trapezius therefore rests on four people, and it is contested because of what happens after the release, which I come to below. A different order of magnitude comes from a measurement of force rather than electrical signal. In seven university archers, in 1993, holding the shooting position to exhaustion corresponded, according to the authors, to about 40-50% of the maximal push-pull force measured on an ergometer, and the hold lasted 52 seconds on average. The claim that holding the shooting position demands between forty and fifty per cent of maximal force is reported, and the two percentages cannot be compared, because one measures force and the other the amplitude of the signal.
The triceps: one muscle, three answers
The most telling case is the triceps of the drawing arm, which is sometimes described as the most loaded muscle in the shot. In a thesis on ten Brazilian national squad archers, 72 arrows each at 70 m and twelve muscles recorded, the triceps was at 73.5% of maximal contraction between fifty milliseconds before and fifty after the clicker fell, the highest value of all, ahead of the finger flexor at 44.1%, the finger extensor at 41.0% and pectoralis major at 40.9%; just after the click it rose to 86.6%. In thirty Japanese archers split into three levels, over the second before the clicker, the same muscle was at 10-20% in each group, and lower in the elite than in the beginners. In fourteen Taiwanese university recurve archers, over the last second before the release, it was at 26% and 34% in the study's two groups, while the upper trapezius and the posterior deltoid were working at between 77% and 95%.
Three studies on archers give three answers, which is why the claim that the triceps of the drawing arm is the most loaded muscle in the shot is contested. The Brazilian study, on national squad archers with many arrows each, cannot be set aside, but its standard deviation for the triceps, 169, is wider than the mean itself, a sign that the value varied enormously from shot to shot or from archer to archer. The difference seems to lie mainly in the method, meaning the time window measured and the way each muscle is related to its own maximum. The Brazilian thesis itself points that way: up to half a second before the click the triceps sat at between 10% and 13%, close to the Japanese figures, and the 73.5% applies only to the tenth of a second around the click.
Shoulders and shoulder blades
If the triceps divides the studies, the muscles around the shoulder blade on the drawing side bring more of them into line. In twenty-four Malaysian compound archers, measured over the whole shot, the shoulder girdle muscles worked on average at between 15% and 32% of maximal contraction; the posterior deltoid and the middle trapezius were the muscles in which most archers held a sustained contraction, 74.6% and 68.6% of the archers according to the authors, and the middle deltoid was the least active. In twenty Polish traditional archers the posterior deltoid gave the largest signal while aiming, but in microvolts and without normalisation. The claim that while aiming, the posterior deltoid and the middle trapezius are among the hardest-working shoulder muscles is reported: neither sample shot recurve, and being among the most active is not the same as doing most of the work alone.
The finding closest to the advice to shoot with your back comes from comparisons between levels. In the thirty Japanese archers, over the second before the clicker, the lower trapezius on the drawing side worked at 54.0% in the elite, 39.4% in the pre-elite and 34.0% in the beginners, while the biceps and triceps worked less in the elite. In twenty-seven Turkish recurve archers the picture was similar, but other muscles separated the levels: the elite used the forearm less and the posterior deltoid and middle trapezius more. This is why the claim that expert archers hold the draw with the shoulder and back rather than the forearm is reported, with two caveats. The part of the back that works hardest changes from study to study, and the forearm never switches off completely: in five elite Russian recurve archers the finger flexor of the string hand worked far harder than its counterpart on the bow arm, with an asymmetry coefficient of 0.86 on a scale running from zero, meaning parity, to one, meaning one side doing all the work.
The bow arm, for its part, is not passive. In eight elite French recurve archers, in the three for whom the value is reported, the middle deltoid of the bow arm worked at between 46% and 66% of maximal contraction while aiming, above every average in the compound sample; as early as 1993, of six recording sites on the deltoid and trapezius of both sides, the bow-side deltoid gave the highest value while aiming. Whether the drawing side works harder than the bow side is something the studies do not agree on: in the five Russian archers the drawing side dominated in most muscles, but serratus anterior, pectoralis major and latissimus dorsi worked harder on the bow side, and in nine recreational archers, with a very light bow, the two sides showed similar levels. The claim that in the shot the drawing side works harder than the bow side is contested.
How much load the shoulder carries in a model
Electromyography percentages do not tell you how much load each muscle carries, because a small muscle at 50% produces less force than a large one at 20%. Estimating that takes musculoskeletal models, which reconstruct forces from movement. For the rotator cuff the only estimate comes from a model built on motion capture of a single adult archer, with a 48 lb bow, during the draw. Read from the graphs, the peaks sit at around 15% of maximal contraction for infraspinatus, 13% for the deltoid, 8% for teres minor and the trapezius, 6% for supraspinatus and below 3% for the rhomboids: the load is shared between the rotator cuff and the deltoid, and no muscle reaches a fifth of its maximum.
These are calculated figures rather than measured ones, from a single archer, and the claim that during the draw the load is shared between the rotator cuff and the deltoid, and no muscle comes close to its maximum is transferred, because it has not been tested on archers. Those values sit well below the 85% of the four Turkish archers and the 77-95% of the Taiwanese ones, partly because the model looks at the draw rather than the aim.
Around the clicker and the release
Muscle activity also changes a great deal in the last second of the shot. The Brazilian study timed it in windows of a hundred milliseconds: eleven of twelve muscles were more active in the fifty milliseconds either side of the clicker falling than in any window up to half a second earlier, and pairs of muscles moved almost in unison, with correlation coefficients between 0.72 and 0.99. The claim that at the clicker drop almost all the arm and trunk muscles are more active than in the half second before, and they move together is reported: it is a single study, and for some muscles, mostly in the trunk, the average over the following hundred milliseconds was higher still, so their peak may come just after the click.
After the click, a lot switches off in the drawing arm. In the four Turkish archers, over the second after the clicker fell, the posterior deltoid dropped from 85.4% to 30.6% and the finger flexors from 83.2% to 33.3%; the extensors fell less, from 82.5% to 57.7%, and the trapezius stayed where it was, at 72.4% before and 72.9% after. In this experiment the picture is the one coaching describes: the fingers stop holding and the back keeps working.
The deltoid, however, behaves in a more complicated way. In 1987 a Japanese group found a brief silence of the deltoid in its best archers: in twelve archers split into three levels, shooting twenty arrows at 8 m, the muscle went quiet in 70-100% of the best group's shots, 110-120 milliseconds after the clicker, while the release came at 170-180; in the other two groups the silence appeared less often. The claim that in expert archers the deltoid falls silent just before the release is reported, because it comes from a single group, with no statistics between levels and no link to the score of individual arrows. In the same work, in the two world-class archers, the deltoid went back to working hard straight afterwards, for about a second and a half, together with the back and shoulders. That is why the claim about 80% and switching off at the release is contested, even though the two measurements concern different parts of the deltoid and the Japanese one reports no percentages.
Experts and beginners: where less work is done
It is said that the expert shoots with minimal effort while the beginner tenses everything. When studies compare elite archers with less accomplished ones while aiming, the muscles of the fingers, arms and trunk almost always work less in the elite: in a French thesis on eight elite and seven intermediate archers at 70 m, eleven of the twelve muscles that differed between the groups were less active in the elite. The twelfth, the middle trapezius of the drawing arm, worked harder, though, and in all three comparisons between levels some part of the trapezius that holds the shoulder blade works harder in the elite. Stated for every muscle, the claim that while aiming, elite archers activate their muscles less than intermediate archers is contested; the reading that holds up best is that the work shifts from the fingers and arms to the shoulder blades.
For the string-hand forearm the picture depends on whether people or arrows are being compared. Between people of different levels the link is there: in twenty young Swedish archers, the tenser the finger extensor in the instant before the release, the lower the score, with a correlation of -0.48 in one of the study's two conditions (in the other, after a session in a flotation tank, the link disappeared), and in twenty-three people measured in Turkey forearm activity around the clicker was related to FITA score with correlations of up to -0.74, although ten of them had never shot and were given an assigned score. Within the same archer, by contrast, in the eight elite French archers, over 549 arrows at 70 m, the string-hand finger extensor was more active on the 10s than on arrows scoring 9 or less. The claim that the less the string-hand forearm muscles work, the better the shot is therefore contested.
Where the research stops
Taken together, these studies answer the first half of the question better than the second. On which muscles switch on, the studies look fairly alike: the posterior deltoid and the trapezius on the drawing side, the deltoid of the bow arm, the forearm muscles that hold the string and, around the clicker, almost everything else at once. How hard they work, as a percentage of their maximum, depends too much on how it is measured to give a single figure: for the same muscle the values run from 10-20% to 73.5%.
Three limits remain. The first is that ranking the muscles would take the same muscles, measured against the same reference test and in the same time window, whereas each study uses its own. The second concerns score: the links between muscle work and the arrows are almost all associations, and exercise programmes for the shoulder-blade muscles have given conflicting results on score, which is why the claim that scapular stabilisation exercises improve score is contested. As for the forearm, nobody has tried teaching one group to relax it to see whether the arrows improve. The third is that the samples are small, often under ten archers per group.
What I take to the shooting line
The first thing I take from this is that the percentages are not a target. No archer can feel whether their deltoid is working at 30% or 80%, and the same muscle gives different figures depending on the instrument and the reference test; using them to tell someone how hard to push or pull has no basis.
The second concerns the image of the drawing arm as a hook. It works as a cue for where to put your attention, but the arm works, and works hard, even in the best archers. So when I look at a draw I care more about how the shoulder blade moves, in a video filmed from behind, than about how soft the hand looks, and I would not ask an archer to switch the forearm off: on the elite French archers' 10s the extensor was more active, not less. How the fingers and the back work at the instant the string leaves is covered in the article on the release.
The third concerns time. In the seven archers of 1993, holding the position took a large share of the available force, and holding it to exhaustion lasted under a minute. A hold like that is far longer than a shot, so the figure does not describe a single arrow, but it does suggest the reserve is not large, which is one reason I also keep an eye on aiming time. For the strength needed to sustain this work, see the article on specific strength, and for the way the muscles coordinate with each other, the one on muscle synergies.