Rowing movements have had a renaissance in recent years. They have become popular for rehabilitation, athletic development and bodybuilding. However, scientific analysis of these movements is still thin on the ground.
Fortunately, there are some recent studies we can turn to for sound information. This one is particularly informative and while some of the conclusions are predictable, some of the results are surprising…
The study: Comparison Of Different Rowing Exercises: Trunk Muscle Activation And Lumbar Spine Motion, Load, And Stiffness, by Fenwick, Brown, and McGill, in Journal of Strength and Conditioning Research, 2009
What’s the background?
At first glance, it’s easy to dismiss the notion that there is any complexity to rowing movements. However, for the enquiring mind, rowing movements present an interesting set of questions. Of course, there are the normal questions that you get with any multi-joint exercise, such as what is the muscle activity during different variations and the relative muscular effort of the various muscles with increasing load.
However, because most rowing movements involve stabilizing the spine, suddenly we have a whole new set of variables to find, including spinal loading, the degree of spinal movement and the spine stiffness. Let’s take a moment to clarify some of these before we delve into the study.
The rowing movement is an antero-posterior one, which means that the direction of force runs perpendicular to the body. This means the forces that are exerted by the weight act at right angles to the spine, creating shear forces. Since shear forces are more damaging to the spine than the compressive forces that occur with axial loading, it is useful to know the shear forces during rowing movements.
However, this isn’t the whole picture, because muscular activity of the spinal muscles actually creates compressive forces, so although the rowing movement is antero-posterior, significant axial compressive forces will also be produced down the spine. And, of course, such shear or compressive forces may be well or poorly tolerated by different populations.
Degree of spinal movement
Since the spine is subjected to significant shear forces during rowing movements, unless the muscles around the spine create forces to prevent its movement, these shear forces will cause flexion of the spine. Additionally, as we will see with single-arm movements, there is also scope for rows to cause rotational movements. And although a certain amount of movement is considered acceptable in most populations, it is often the case that reduced spinal movement is better for many individuals.
Spine stiffness describes the ability of the spine to remain motionless in the sagittal plane while being exposed to large shear forces. If the spine moves little during high loads, this would be considered very stiff. This stiffness of the spine under load is governed by the activity levels of the muscles that surround it. So we might expect a reasonable correlation between compressive forces and stiffness. Also, spine stiffness is generally considered a very good proxy for spinal stability, or the ability of the spine to maintain its position during perturbations.
What did the researchers do?
The researchers wanted to investigate three different rowing exercises: the bent-over barbell row, the inverted row and the one-arm cable row. In doing so, they set out to establish the muscular activity levels of the back, leg and abdominal muscles, the spinal loading, the degree of spinal movement and the spine stiffness. So they recruited 7 young and healthy, recreationally active men to perform the three different exercises.
Muscle activity measurements
While the subjects performed each lift, the researchers monitored the electromyographical (EMG) activity of various leg, back and abdominal muscles, including the right and left rectus abdominis, right and left external oblique, right and left internal obliques, right and left latissimus dorsi, right and left upper (thoracic) erector spinae, right and left lumbar erector spinae, right gluteus medius, right gluteus maximus and right biceps femoris.
Spinal movements, loading and stiffness
The researchers used video cameras to monitor the movements of the spine during the study. They entered all of the information, including the normalized EMG activity and the movement data from the video into a previously-created spinal model, which estimated the forces acting on each vertebra and also the stiffness.
The researchers recorded the muscle activity of the various muscles and reported them as a percentage of MVC. However, it was not clear from the study whether these values were mean or peak. The following chart shows the leg muscle activity reported in the study:
The chart shows that the various rowing movements produced very different levels of leg muscular activity. The one-armed cable row and the bent-over row produced much more rectus femoris (hip flexor/knee extensor) activity than the inverted row. However, the inverted row produced much more biceps femoris (hip extensor/knee flexor) activity than the other two exercises.
Similarly, the inverted row and bent-over row produced much greater gluteus maximus activity than the one-armed cable row. And the inverted row produced more gluteus medius activity than the other two exercises, particularly the one-armed cable row.
The following chart shows the back muscle activity of the various rowing movements:
The chart shows that the one-armed cable row has a very different pattern from the other exercises in terms of back muscle activity. The one-armed cable row (with the right arm) produced much less back muscle activity in the right and left lower erector spinae and very little (obviously) in the left upper erector spinae and left latissimus dorsi. However, this is because of the one-arm nature of the row.
Perhaps more interesting is that the latissimus activity is clearly higher in the right arm of the one-arm cable row and in the inverted row than in the bent-over barbell row. Similarly, the inverted row displays a tendency to produce more upper erector spinae activity than the bent-over row but less lower erector spinae activity than the bent-over row.
The following chart shows the activity of the various abdominal muscles:
The researchers noted that the one-arm cable row demonstrated significantly higher activity of the right internal oblique and the left external oblique, as might be expected during the one-arm movement using the right arm. Other than these differences, the abdominal activity during the various rowing movements appears surprisingly similar.
The researchers found that spinal movements were very different between the three rowing variations. The two key findings were (1) that the bent-over row produced much greater flexion/extension movement than the inverted row or the one-armed cable row, and (2) that the one-armed cable row produced a very high degree of spinal rotation. The following chart shows the findings:
The chart shows the spinal movements in a normalized format, so as a percentage of the maximum available range of motion in the spine. So you can see that the bent-over row uses c. 40% of the maximum available flexion/extension motion, while the one-armed cable row uses nearly 70% of the maximum available rotational motion.
Comparison of spinal loading and stiffness
The researchers reported the compressive, anterior-posterior and medial/lateral shear forces acting on the L4/L5 lumbar vertebra. The results are shown in the following chart:
The researchers found that the standing bent-over row displayed very significantly higher compressive forces as a result of the muscles contracting around the spine. However, there were no significant differences between anterior-posterior and medial/lateral shear forces, which were found to be very small.
The researchers found that stiffness of the spine also varied between exercises, as shown in the chart below:
The chart shows that the flexor stiffness was greatest in the bent-over row and lowest in the one-arm cable row. However, axial stiffness was similar in all exercises.
What did the researchers conclude?
The bent-over row produced the lowest levels of latissimus activity, high levels of upper erector spinae activity and the highest levels of lumbar spine activity, suggesting that the lower back muscles were being worked significantly during this exercise.
The researchers concluded that bent-over rows produced the greatest amount of flexion/extension movement of the lumbar spine. Bent-over rows were also found to produce the highest amount of spinal compressive forces at L4/L5 and led to much greater values of spine stiffness. The researchers suggested that the large antero-posterior forces produced a large moment on the lumbar spine, which the lower back muscles had to resist, thereby producing very large compressive forces and stiffness to stabilize the lumbar spine.
Summary: overall, the bent over row is an evenly-distributed upper-and-lower back exercise but produces high compressive forces, which tests the ability of the spine to stabilize itself.
Use for: developing the upper and lower back evenly in individuals without lower back problems.
The inverted row produced the highest levels of latissimus and upper erector spinae activity but lower levels of lower erector spinae activity. It produced very little rectus femoris (hip flexor-knee extensor) activity but high biceps femoris (hamstring) activity. The low hip flexor activity may be useful for those looking to reduce anterior pelvic tilt. It also produced the highest levels of activity in the gluteus medius but low levels in most of the abdominals.
The researchers noted that the inverted row produced much higher upper back activation than lower back activation, which they noted would be ideal for an early rehabilitation program in which lumbar loading was contraindicated. However, they also thought that it might lead to an imbalanced level of development in an advanced rehabilitation setting. However, in an athletic development setting, where training the upper back is often required with greater frequency than the lower back, an exercise that develops the upper back while not stressing the lower back is very valuable.
The researchers concluded that inverted rows produce the least overall movement of the spine. They suggested that inverted rows might therefore be the best option for those with an overall intolerance to spinal movement. However, inverted rows do still produce nearly the same amount of flexion/extension movement as bent-over rows.
Summary: the inverted row is the most effective of the three exercises for targeting the lats and the upper back while simultaneously sparing the spine, avoiding stress on the lower back and avoiding recruitment of the hip flexors.
Use for: developing the upper back and lats specifically in individuals with lower back problems or in circumstances where additional upper back/lat development is required.
One-arm cable row
The one-arm cable row also produced high levels of latissimus and upper erector spinae activity but the lowest levels of lower erector spinae activity. It produced very little biceps femoris (hamstring) activity but high rectus femoris (hip flexor-knee extensor) activity. The high level of rectus femoris activity may be unhelpful for those rehabilitation settings in which excessive anterior pelvic tilt is present. It also produced the highest levels of activity in the external and internal obliques.
The researchers concluded that one-arm cable rows produced the greatest amount of rotational motion of the spine, which is nearly 70% of its maximum available motion. This suggests that for individuals who are rotational motion-intolerant, the cable row might be a poor choice. The researchers noted that this movement was very small in absolute terms but very large in relative terms, so it could easily be missed, making the one-arm cable row an exercise that should be carefully monitored for form.
Summary: the one arm cable row is another upper-back and latissimus-focused exercise that activates the obliques significantly. Its only downfall is that it produces large relative rotational movements, which may be unhelpful for rehabilitation settings.
Use for: developing the upper back and lats in conjunction with the obliques. Make sure to monitor form carefully to ensure no rotational movement of the spine.
What were the limitations?
One of the key limitations to this study was that some key mid-back muscles weren’t tested, such as the mid-traps and rhomboids. It would be interesting to see how the exercises compared in terms of scapular retractor activation.
Another obvious limitation was that the chest-supported row wasn’t covered, as this exercise has a support for the torso, which may have reduced the spinal movements and forces significantly, making it an even better choice for rehabilitation than the inverted row.
What were the key points?
The key points were as follows:
- For targeting the lats and upper back, the inverted row is superior to the bent-over row. However, the bent-over row produces greater overall back muscle activity, including the lower back, and challenges spinal stability the most.
- The bent-over row creates large compressive forces and flexion/extension movements, while the one-arm row produces large rotational movements. The inverted row does not cause the same relative movements or involve the same forces. The inverted row may therefore be better suited for those with lower back trouble.
What are the practical implications?
For everyone strength training:
To make sure you can train well into old age, build your upper back and latissimus routine around weighted inverted rows rather than bent-over rows or one-arm cable rows and avoid both spinal forces and rotational movements.
Want more research reviews?
If you liked this post and you haven’t already subscribed to our monthly review service, you can buy a subscription here. If you’re not ready for the paid service yet, you can sign up to our free newsletter here.