What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Surface Tension - Aero Drag

Turbo911

Member
Back in the day, I raced RC boats. We used Imron paint over carbon/kevlar laminates to finish the boat. The first time I took my bright glossy boat to the lake, it would not get up on plane. This was a proven design that would go 100 mph. I was stumped and started looking at engine and drive train. Then someone came up to me and said "Your boat will never get up on plane until you scuffed the bottom of the boat". I said "I just spent days prepping and painting this thing to get it perfect." He replied "Well, it will make a nice looking bouy until you scuff the bottom of the boat.
Eventually, I gave in. The boat immediately took off on plane. I went home and did some research. He was absolutely correct, the shiny paint finish on the bottom of the boat was allowing the water to just grap the boat with surface tension.
My point to all this is, does this apply to air over a surface?
Eventually, I scuffed the top of the boat as well to reduce aero drag and picked up 8 more MPH.
Interest to hear if anyone has experimented with this? The closest thing might be a painted plain vs an unpainted plane. As most do not paint until the 40 hours are flown.
 
Data

I would be interested in seeing data on the extra 8 mph by scuffing the upper surface of a model boat.

The glass smooth bottom, however, is definitely an issue. I had the same experience...
 
The aerodynamic for aircraft is different because of the surface tension you mention.

Here a Youtube link that explain why the AR5, the world speed record holder aircraft, flew so fast. The link explains the important of getting the aircraft skin surface smooth and uniform.

https://youtu.be/rxvoDbZpoY8?t=1983
 
Reminds me of a buddy that refused to get his Bonanza repaired after it got pelted with hail. He claimed it was faster with the golf ball effect applied.
 
Quite a few studies performed back in the 80’s with “riblets”. Worked on high speed aircraft as well as low speed watercraft. Similar Reynold’s numbers possibly?

https://apnews.com/7593cedd8df4e6d218f3aae9bfdacb14

At what airspeeds is this effective? I found a reference to Cathay Pacific operating A340s with riblet film (not clear how much of the airframe has it applied) with a demonstrated reduction of skin friction drag of 5%-8%. I would imagine that the drag reduction at lower airspeeds would be somewhat less. I further assume that it's proportional to TAS, not IAS, but I could be wildly wrong. Any insights from the aero engineering crowd?

https://www.sciencedirect.com/topics/engineering/riblets
 
Reminds me of a buddy that refused to get his Bonanza repaired after it got pelted with hail. He claimed it was faster with the golf ball effect applied.

I had a Piper Turbo Arrow III many years ago that had hail damage all over it. The plane was faster than any other Turbo Arrow I've flown. We figured it might be the golf ball effect.
 
I had a Piper Turbo Arrow III many years ago that had hail damage all over it. The plane was faster than any other Turbo Arrow I've flown. We figured it might be the golf ball effect.

OK, just who is going to be the first to scuff their paint job and add the golf ball dimples to test the speed? If roughness increases speed why don't universal rivets go faster than countersunk?
 
Last edited:
To the OP and others, gasses do not have surface tension like liquids do. So in air, there is no surface tension problem.

Riblets work in the micro sublayer of the turbulent boundary layer. NOT for laminar flow. So where you put them really matters. Orientation of the riblets is crucial. You can not reproduce the riblet effect by sanding - the profile shape of the scratches from sanding is not the right shape.

Flight tests on jet liners have confirmed the benefit of riblets, but operational problems negate the benefit. In particular, the plastic film is not porous enough to breath out the minute leak at each rivet of a pressurized fuselage, so the film blisters at each rivet. Cleaning and maintaining the riblet film is also difficult to do properly.

One of the most significant uses of riblets was on the hull of the America's Cup boat Stars and Stripes, which raced against Australia's Kookaburra III in Freemantle and reclaimed the cup.

For general aviation airplanes, you want the surface as smooth as you can where ever there is laminar flow (first 15--20% chord on RV 3,4,6,7,8, back to the spar on the RV-10 and RV-14 if the riveting is done well, and a good fraction of the lower surface of the RV-9 if the riveting is done well) I have not been able to measure any increase in top speed on my RV-8 with wings clean compared to covered with bugs. I imagine that the -10 and -14 should demonstrate a measurable benefit to clean smooth wings fwd of the spar.

Once you have turbulent boundary layers, it really doesn't matter much, as long as the surface is not outright "rough".
 
Last edited:
So in actual fact my buddy may have been wishful he was faster with the pitted surface? He was known to exaggerate his golf scores too.

Steve, thanks for your insights.
 
So in actual fact my buddy may have been wishful he was faster with the pitted surface? He was known to exaggerate his golf scores too.

Steve, thanks for your insights.

The dimples on golf balls improve performance in two ways. Neither involves reducing skin friction drag.

First, roughness causes earlier transition to a turbulent boundary layer, which stays attached longer around the adverse pressure gradient to the back side of the ball, so the separated-flow wake is smaller. This is a reduction in pressure drag which outweighs the increase in skin friction drag. Most any roughness would do this, don't need dimples. But no relevance to shapes that have little or no pressure drag from flow separation.

Second, the dimples drag the boundary layer along with the spinning ball, causing circulation, which is lift. A golf ball in flight has a significant amount of lift that extends its flight. The stitching seams on a baseball or cricket ball do the same thing, allowing the flight path to be curved in the direction of lift. (An exception is the "reverse swing" in cricket, which is a somewhat complex interaction of both effects described here.)

There have been balls developed that have different dimple patterns in different areas to optimize for these two separate effects, and perform significantly better. But they are illegal in regulation golf. The dimple pattern must be symmetrical/uniform.

If you want to put a Lycoming on a giant sphere and sit in it while it has lots of backspin, by all means put dimples on it.
 
Last edited:
After getting my old books out and doing a little more research. The surface of a wing shape needs to smooth and/or straight as to not interrupt laminar flow. Waviness, bugs, dents and rivet heads all would cause the disruption of laminar flow. I am specifically talking about the surface boundary of the shape going thru the air. There is some data that suggest a that has been sanded (like wet sanded) can produce less drag than perfect painted surface.

The surface boundary layer is a thin (a few thousands of an inch to maybe 0.03" of an inch of air. My belief is that the friction of air/air is less than air/painted surface.

Hopefully a fluid dynamics person will chime in.
 
Last edited:
This is a well known effect that is used in racing sailboats and has been explored in gliders. There is a video of an old aerodynamics course somewhere online where this is demonstrated in a lab. They use a smooth ball, measure the drag, then abrade it with a specific grit paper in a direction perpendicular to the flow. When sized properly, the ridges and grooves of the abraded surface serve to energize the boundary layer to promote laminar flow over larger areas of the wing.

The problem is that the size of the ridges and grooves is pretty sensitive to get the desired effect. In air, things like dirt and bugs can have adverse effects, cancelling any gains.
 
thew truth and the myth

Boats move over the surface of a media (water), while aircraft move through a media (air). Back in the day, some old salty pilots used to climb a bit above the cruise altitude, and descend to pick up a bit of speed, and called that "putting it on the step". Many a co pilot got chewed out by a captain for letting the aircraft fall "off the step". BS!!! An aircraft will always cruise at its trimmed speed for the power and drag. (I.E. in stable conditions, fuel flow will create a fixed angle of attack, and a fixed cruise speed until something disturbs or changes the circumstances. Its not a boat, its an aircraft (airplane is a bit misleading, they do not plane!!!).
Cheers!

Gary
 
I remember the story that when the Rutan Vari-Ezes hit the scene many of the builders complained about performance and speeds vs factory specs. Burt always built his prototypes quick and dirty without a lot of perfection profiling. One particularly persistent builder with and amazing show-quality finish kept giving Burt the red *** about it at Oshkosh. So Burt finally make a deal with the guy in front of witnesses that if he could get the fellow's airplane up to spec before the end of Oshkosh the builder would agree to forever shut up. As soon as the deal was shook upon in front of everybody involved Burt walked over the airplane and started sanding the canard with 60 grit sand paper alternating 45 degree angles. The builder had a coronary but was reminded a deal was a deal. After Burt sanded the canard and wing until he figured the surface was to his suiting he turned to the builder and said, "There you go. Now go fly it". Part of the Eze lore.
 
This is all false on two counts. First of all, adding roughness to a ball reduces drag by reducing the pressure drag, not the skin friction drag. See discussion above about golf balls. Airplanes have very little pressure drag, so this is not relevant. What changes the drag on a sphere has no relation at all to what changes the drag on a wing or fuselage.

Further, ridges, grooves, and other roughness does not promote laminar flow, they destroy it. This happens because the minute disturbances are amplified in the boundary layer until they become unstable and cause transition to a turbulent boundary layer. For reference, look up Tolmien-Schlichting instability.

Minute scratches and grooves, if done exactly right, can reduce the skin friction in a turbulent boundary layer. This is what riblets do. The grooves inhibit the minute mixing that transports momentum out of the micro sublayer of the boundary layer. The grooves should be parallel to the flow.

Another thing that is getting confused in here is the effect of careful sanding on glider wings. This is not done to promote laminar flow of the clean, dry wing. The scratches promote the wetting out of water droplets when we have the bad luck of having to fly in virga or rain. Water droplets have a disastrous effect on laminar flow, and fine sanding scratches help the droplets wet out smoothly, causing far less disruption of the laminar flow, although still causing some early transition. This is a trade-off where a minute performance loss in dry conditions is accepted in exchange for a substantial improvement of performance when rain drops are present. Except for the first few inches of airfoil, which really should be polished smooth, the fine sanding scratches (800 grit or so) are not really big enough to seriously reduce the amount of laminar flow. But in no way do they promote it either!

This is a well known effect that is used in racing sailboats and has been explored in gliders. There is a video of an old aerodynamics course somewhere online where this is demonstrated in a lab. They use a smooth ball, measure the drag, then abrade it with a specific grit paper in a direction perpendicular to the flow. When sized properly, the ridges and grooves of the abraded surface serve to energize the boundary layer to promote laminar flow over larger areas of the wing.

The problem is that the size of the ridges and grooves is pretty sensitive to get the desired effect. In air, things like dirt and bugs can have adverse effects, cancelling any gains.
 
This is all false on two counts. First of all, adding roughness to a ball reduces drag by reducing the pressure drag, not the skin friction drag. See discussion above about golf balls. Airplanes have very little pressure drag, so this is not relevant. What changes the drag on a sphere has no relation at all to what changes the drag on a wing or fuselage.

Further, ridges, grooves, and other roughness does not promote laminar flow, they destroy it. This happens because the minute disturbances are amplified in the boundary layer until they become unstable and cause transition to a turbulent boundary layer. For reference, look up Tolmien-Schlichting instability.

Minute scratches and grooves, if done exactly right, can reduce the skin friction in a turbulent boundary layer. This is what riblets do. The grooves inhibit the minute mixing that transports momentum out of the micro sublayer of the boundary layer. The grooves should be parallel to the flow.

Another thing that is getting confused in here is the effect of careful sanding on glider wings. This is not done to promote laminar flow of the clean, dry wing. The scratches promote the wetting out of water droplets when we have the bad luck of having to fly in virga or rain. Water droplets have a disastrous effect on laminar flow, and fine sanding scratches help the droplets wet out smoothly, causing far less disruption of the laminar flow, although still causing some early transition. This is a trade-off where a minute performance loss in dry conditions is accepted in exchange for a substantial improvement of performance when rain drops are present. Except for the first few inches of airfoil, which really should be polished smooth, the fine sanding scratches (800 grit or so) are not really big enough to seriously reduce the amount of laminar flow. But in no way do they promote it either!

In a lay man's terms, no need to scuff the paint?
 
And no rv is going to have significant laminar flow. Too many rivets. If you want to reduce drag, fair in all control horns, make your cowl and intersection fairings fit perfectly. Fair in anything like sump drains, tank vents, anything that will generate a wake. That wake is the speed killer. There are lots of reasonably free knots to be had with a bit of work. The first few are fairly easy. They get progressively harder and more expensive the more you try to get.
 
I would be interested in seeing data on the extra 8 mph by scuffing the upper surface of a model boat.

The glass smooth bottom, however, is definitely an issue. I had the same experience...

Sorry, this was approx 20-25 years ago, although we did have a radar gun. This was un scientific as we did tests on different days. However the second day when we picked up the speed the air density was lower by 2%, so we attributed the gain in aero to the top of the hull.
 
And no rv is going to have significant laminar flow. Too many rivets. If you want to reduce drag, fair in all control horns, make your cowl and intersection fairings fit perfectly. Fair in anything like sump drains, tank vents, anything that will generate a wake. That wake is the speed killer. There are lots of reasonably free knots to be had with a bit of work. The first few are fairly easy. They get progressively harder and more expensive the more you try to get.

This is true and solid advice, except to say that when I designed the airfoil for the RV-10 (and now also on the -14) it was designed to achieve laminar flow back to the spar seam. If you do a really good job with rivets, you have a reasonable chance of keeping the laminar flow even along the rivet rows, for a portion of the leading edge. Likely not all the way to the spar. This also depends on being smooth across the spanwise stiffeners (which I did not assume would be there ;) )

So PLEASE try not to paint any stripes on the forward portion of your RV-10 or RV-14. Unless you sand to blend the stripe edge perfectly, you will kill any hope of laminar flow.

Also, when John Roncz designed the airfoil for the RV-9, it embodies a shape that would achieve a very large run of laminar flow on the lower surface. With a skin seam at the spar, no chance beyond that, but again, if your rivets are really flush and smooth, you will get a fair bit of laminar flow on the -9.
 
Last edited:
This is all false on two counts. First of all, adding roughness to a ball reduces drag by reducing the pressure drag, not the skin friction drag. See discussion above about golf balls. Airplanes have very little pressure drag, so this is not relevant. What changes the drag on a sphere has no relation at all to what changes the drag on a wing or fuselage.

Further, ridges, grooves, and other roughness does not promote laminar flow, they destroy it. This happens because the minute disturbances are amplified in the boundary layer until they become unstable and cause transition to a turbulent boundary layer. For reference, look up Tolmien-Schlichting instability.

Minute scratches and grooves, if done exactly right, can reduce the skin friction in a turbulent boundary layer. This is what riblets do. The grooves inhibit the minute mixing that transports momentum out of the micro sublayer of the boundary layer. The grooves should be parallel to the flow.

Another thing that is getting confused in here is the effect of careful sanding on glider wings. This is not done to promote laminar flow of the clean, dry wing. The scratches promote the wetting out of water droplets when we have the bad luck of having to fly in virga or rain. Water droplets have a disastrous effect on laminar flow, and fine sanding scratches help the droplets wet out smoothly, causing far less disruption of the laminar flow, although still causing some early transition. This is a trade-off where a minute performance loss in dry conditions is accepted in exchange for a substantial improvement of performance when rain drops are present. Except for the first few inches of airfoil, which really should be polished smooth, the fine sanding scratches (800 grit or so) are not really big enough to seriously reduce the amount of laminar flow. But in no way do they promote it either!

Steve,

Good luck!:rolleyes:
 
From fact to fiction!

It is amazing how the details of a story can morph in folk lore from something that is probably true to something that is completely backwards!

Rutan well-understands the importance of smooth, wave-free contours on highly loaded canard airfoils.

If Burt ever sanded anyone's canard, it would have been with 600 grit, not 60 grit!

Canard airplanes (that are statically stable) ask a lot out of a canard. It is a very highly loaded lifting surface. To achieve the high lift values with good performance, airfoils are used which require good laminar flow on the forward portion, followed by good transition to turbulent boundary layer to survive the aggressive adverse pressure gradient on the rear portion of the airfoil.

The laminar portion must be smooth and free of waves. Sanding would be a must to remove any waviness. Even orange peel from unsanded paint may be enough to cause early transition, which will kill the lift and increase drag, both. Very fine sanding (at 45 degrees is good) will help with wetting out rain drops that can seriously degrade the lifting ability. Otherwise, after sanding to remove any waviness, polished would be best.

Canard lore is full of stories of rain causing serious problems, and especially a painted accent color stripe along the leading edge. The step along the masking line between colors would be a disaster, could even prevent a canard airplane from rotating for takeoff.

Now, you might reasonably say, "Hey, wait a minute, you keep telling us that a turbulent boundary layer can better survive adverse pressure gradients without separation (such as roughness on a ball) and yet you are telling us that early transition to turbulent boundary layer caused by bugs, rain, waviness, or lack of smoothness will kill the lift. This seems contradictory!"

And it does seem so. But here is the distinction. In order for a turbulent boundary layer to survive the very aggressive adverse pressure gradient on a thick, highly cambered airfoil (look up Liebeck airfoils), the turbulent boundary layer must be fresh, young, and healthy. Well energized. If you have premature transition because of bugs, rain, roughness on the forward portion of those airfoils, the turbulent boundary layer is pretty tired by the time it gets back to where the adverse gradient starts, and so, even though turbulent, it will separate. This results in significant lift loss (stall) well below the angle of attack where stall would normally occur on the same airfoil that is properly smooth. And of course, even at angles below that premature stall angle, there is also a significant increase in drag from the loss of laminar flow.

You might also point out that there are times when forcing transition from laminar to turbulent boundary layer is a good thing. We use zig-zag tape, or bumps, or blow-holes, or a variety of other tricks, to purposely cause transition just ahead of some place where a fresh, young turbulent boundary layer is needed to reduce/prevent separation. Often just ahead of control surface hinge lines, or on the underside of a aft-loaded undercambered airfoil. On small model airplanes that fly at VERY low Reynolds number, it is often beneficial to try to trip the laminar boundary layer right at the leading edge, or even from a trip wire suspended out in front of the wing. Actual transition will not occur until some point downstream, as the T-S waves amplify enough to cause transition. In all these cases, the higher drag of the turbulent boundary layer is accepted as a consequence of achieving better flow attachment downstream.

I remember the story that when the Rutan Vari-Ezes hit the scene many of the builders complained about performance and speeds vs factory specs. Burt always built his prototypes quick and dirty without a lot of perfection profiling. One particularly persistent builder with and amazing show-quality finish kept giving Burt the red *** about it at Oshkosh. So Burt finally make a deal with the guy in front of witnesses that if he could get the fellow's airplane up to spec before the end of Oshkosh the builder would agree to forever shut up. As soon as the deal was shook upon in front of everybody involved Burt walked over the airplane and started sanding the canard with 60 grit sand paper alternating 45 degree angles. The builder had a coronary but was reminded a deal was a deal. After Burt sanded the canard and wing until he figured the surface was to his suiting he turned to the builder and said, "There you go. Now go fly it". Part of the Eze lore.
 
Last edited:
Steve, thanks for this excellent discussion of forms of drag. As a teacher of functional morphology of fish and birds, you have clarified a couple of points for me. And thanks for that awesome wing on the RV-10, it continues to impress. And FWIW, I do see a bugs on the wing speed penalty.
 
Take aways

Hey guys and gals, great conversation. My take away from this thread is that the analogy of scuffing the bottom of my R/C boat really does not apply to aircraft. Kind of apples and oranges. I really appreciate all the comments and the technical expertise that was shared here. Truly informative and got me to do some more research that I had long since forgotten.
 
CAFE study

I recall reading a CAFE study where a Mooney or a Bonanza was flown through a rather specific set of tests, instrumentation and such, then attacked by a small army of volunteers for waxing and reflown within a short period of time. I remember being impressed by the effort and the favorable results. My Google-fu is off and I cannot find that study at the moment.

I wax mine just because the bugs come off easier. :)
 
water is different than air primarily because of water surface tension property.... and you can drink water.
 
So the obvious (to me) question is whether/how this applies to the prop airfoil? Presumably a polished prop would be more efficient than a buggy and scratched one?

And secondly, with the issue of rivet heads, would it make sense to wrap the leading edges of our airfoils to cover (i.e., smooth out) any irregularities caused by less-than-perfect riveting?
 
So the obvious (to me) question is whether/how this applies to the prop airfoil? Presumably a polished prop would be more efficient than a buggy and scratched one?

And secondly, with the issue of rivet heads, would it make sense to wrap the leading edges of our airfoils to cover (i.e., smooth out) any irregularities caused by less-than-perfect riveting?

It depends on the prop. I have heard several owners report a decrease in prop performance on Catto prop if they apply a plastic protective film, presumably because of the step at the edge of the plastic film. And a few of us imagine our WW 200RV prop pulls better when the nickel leading edge is kept smooth -- but that is not measured, just a sense.

If it were just rivet head edges, a plastic film might help. The real issue is the shallow but large crater around each rivet. Even if almost inperceptable, it may be enough to cause premature transition. There was a study by Dan Somers at NASA Langley years ago, that compared metal laminar-flow wing sections with paint, waxed or not, and plastic film, or bare.
 
I gained 3kts after removing the prop tape on my catto. It was the cheapest and easiest knots I found on the airframe.
 
Steve said: "The real issue is the shallow but large crater around each rivet. "

Do you refer to the shallow dip or depression that occurs, for example, when "gun riveting" skin to a rib or formed bulkhead?

I was taught to "recover" from or "restore level" from the dip by backing the formed head with a bucking bar and lightly and "squarely" tapping the skin with a "fresh" rubber mallet.
 
Steve said: "The real issue is the shallow but large crater around each rivet. "

Do you refer to the shallow dip or depression that occurs, for example, when "gun riveting" skin to a rib or formed bulkhead?

I was taught to "recover" from or "restore level" from the dip by backing the formed head with a bucking bar and lightly and "squarely" tapping the skin with a "fresh" rubber mallet.

That is what I am referring to. Once the skin is locally yielded in the driving process, I don't know if you can fully remove it. Better technique certainly helps. I have certainly not been taught the finer points.

I know of some Rocket builders that perfected a process for solution treating AD rivets so they were soft and took less driving force, then allowed them to re-age harden over time, recovering full strength. The lower driving force of the soft rivets allows much smoother surfaces.
 
I've seen a couple airplanes built that way. They are remarkably smooth. But the ones I saw were also glued and then riveted, so it might be a combination of techniques. They were built by a meticulous and careful builder.

Dave
 
Technique

I am no expert, but I did stay at a Hotel last night.

I found the smoothness of the rivet can be controlled by how hard the gun is pressed against the rivet, verses how hard the bucking bar is pressed to form the shop head. I also found this seems to be dependent on air pressure/flow when driving. JMHO.

I worked with great care, when doing the wings, to keep the surface as smooth as possible.
 
Reminds me of a buddy that refused to get his Bonanza repaired after it got pelted with hail. He claimed it was faster with the golf ball effect applied.

Years ago back when I flew for a commuter airline we had one aircraft that went through a hail shaft. Anything facing the relative wind was just destroyed. It was a fairly new aircraft so they put windshields in it, speed taped the leading edges and took it back to the factory (on a ferry permit of course) where they re-skinned it. When it came back it was never quite the same. If you made any power/speed changes you had to re-trim all three axis. However, based on indicated airspeed it was the fasted ship in the fleet by between 5 and 10 knots.
Conceptually we all knew it was some sort of instrument error because the leg times were exactly same as any other aircraft in the fleet but it was satisfying to be going "faster".
 
With the caveat that no two RV's are alike (just like factory airplanes), I have an observation. I regularly fly with an almost equivalently-equipped RV-8 (including this morning to TorC New Mexico). We both have 180hp IO-360's with the Hartzell blended airfoil prop and G3X Touch panels. They are so similar we can jump between them and feel right at home right down to the systems. He spent 11 years building a slow-build kit with Grove gear and the builder of my 8 spent 12 years carefully building a fast-build kit with factory gear. Both have low-profile Todd's canopies and rocket-link steering.

The only real difference between between the two airframes is his aircraft finish was filled and sanded smooth prior to prime and paint. So most of the rivets and even fuel tank seams are invisible. I don't particularly like that idea on a flexible aluminum airframe but he went with it and it has held up well for 500 hours and 5 years. He used a specialized filler that is alleged to adhere and flex with the surface. I sure hope it hangs in there because it's an absolutely beautiful airplane. I have been thinking it's his smooth profiled wing surface that gives him the faster speed but I don't know for sure. I did read Schlichting's Bounday Layer Theory text when taking a course on it back in engineering school (Gig 'Em). But that was 30 years ago. My airplane just might be a little bent somewhere.

Jim
 
Last edited:
For what its worth I owned and flew a C85 powered hand-proppped Vari-Eze back in the early 90's. When in any kind of precipitation the canard would loose lift requiring more back stick to keep it in the air. Alarming, especially at first. When this occurred the pilot had the perfect view of the rain drops stopping on the top surface of the canard in a collective standing wave and right behind it the aft third of the canard surface would be dry. In fact if any moisture found it's way onto the aft third it would reverse direction and travel forward on the surface. It was like the water tripped the boundary layer to the point of separation and localized reverse flow. Interesting to watch while trying not to run out of nose up trim. I think this problem was why the Roncz canard was later developed.
 
That sounds exactly like a laminar separation bubble. Usually more of a thing as Re drops.

348_2015_2031_Fig1_HTML.gif
 
Back
Top