Where and How could the Wallace Experiment Easily Be Repeated?

Mick West

Administrator
Staff member
A classic demonstration of the curvature of the Earth was performed by Alfred Russel Wallace in 1870:

http://blogs.scientificamerican.com...modern-biology-got-suckered-by-flat-earthers/
The iron parapet of Welney bridge was thirteen feet three inches above the water of the canal. The Old Bedford bridge, about six miles off, was of brick and somewhat higher. On this bridge I fixed a large sheet of white calico, six feet long and three feet deep, with a thick black band along the centre, the lower edge of which was the same height from the water as the parapet of Welney bridge; so that the centre of it would be as high as the line of sight of the large six-inch telescope I had brought with me. At the centre point, about three miles from each bridge, I fixed up a long pole with two red discs on it, the upper one having its centre the same height above the water as the centre of the black band and of the telescope, while the second disc was four feet lower down. It is evident that if the surface of the water is a perfectly straight line for the six miles, then the three objects—the telescope, the top disc, and the black band—being all exactly the same height above the water, the disc would be seen in the telescope projected upon the black band; whereas, if the six-mile surface of the water is convexly curved, then the top disc would appear to be decidedly higher than the black band, the amount due to the known size of the earth being five feet eight inches, which amount will be reduced a little by refraction to perhaps about five feet.
Content from External Source
1905_Wallace_A237.2_fig399.jpg
(Original source is Wallace's "My Life", Page 364-376)


Put more simply, we have:

A) A telescope (which we would replace with a digital camera with powerful zoom lens)
B) A target 3 miles away
C) A target 6 miles away, in line with A and B.

A,B, and C are all the same height above the waterline. So you just look from A towards B and C, and if B is visually higher than C then the water surface is curved up in the middle.

Rather surprisingly I can't find any well documented recreations of this experiment. It seems like it would be an ideal science project. There are significant difficulties though as you need a suitable stretch of water, the targets are 12 feet above the water, and the six mile distance makes it difficult to coordinate.

But it's really a very straightforward demonstration. So what I'd like to do in this thread is discuss places where the experiment might best be carried out, and what procedures could be used to make the execution of the experiment as straightforward and foolproof as possible.
 
Last edited:
One idea to open up more locations would be to use a periscope at point A. You could make quite tall lightweight periscope with very simple cheap construction This gives you a lot more flexibility in terms on positioning the camera and targets.

You could use some 1" PVC Schedule 40, which would also be good for the target supports. You could make a clip-together or telescoping arrangement of the PVC pipe for easy transport, even adjustable heights. It's $4 for a 10' section.
 
A series of three bridges such as in the original seems ideal, as you can use a simple plumb measure to get the height.

The original Bedford Level is still there, and has been used to kind of recreate the useless "man in a boat" version of the experiment. But it's not clear if it's still suitable for the Wallace experiment. These seem to be the bridges:
20160912-141904-gg5f7.jpg (See attached Google Earth File)

However the middle bridge is a railway bridge and does not seem to have access.
 

Attachments

  • Bedford Level.kmz
    1.1 KB · Views: 745
A flowing body of water will have a sloped surface. Else it wouldn't be flowing. There has to be some factor figured in for that.
 
A really long straight causeway could be a opportunity for multiple targets. Might be tricky though

Seven Mile Bridge, Florida



Pontchartrain causeway, Lousiana
 
I live next to this lake. There are 4 possible places to do this, tell me what you think:









The locations of the left hand side from the pictures, from top->down, are:

-A dock/boat loading area
-Beach/boat loading area
-Beach/boat loading area
-A small water inlet with a boat loading area

The right side locations from the pictures, from top->down, are:

-Accessible shore from road, or by boat
-Accessible shore from road, or by boat
-Accessible shore from road/park, or by boat
-Accessible shore from road/park, or by boat

Also, many people ice fish on this lake in the winter.. (and would probably ice fish in the summer if there was ice! Wisconsin!!!)
 
It's hard to get a spot exactly in the middle. At this scale we can approximate the curve as a triangle, and for an offset midpoint at a distance m from the end (where m<d/2), then we can approximate the "bulge" as bulge*2m/d

So for 5 miles, bulge is 4.17 feet, but if the middle marker were at 1.5 miles, then it 4.17*2*1.5/5 = ~2.5 feet.

So a 5 mile span should still work, even if you can't get the middle exactly.
 
Last edited:
Folsom Lake is plausible. There's a 4 mile stretch with a pretty centered spit of land (depending on the water level)20160912-171849-9vwfr.jpg

The middle is only really accessible by boat though.
 
It's hard measuring from a water surface in a canal, etc. But if you are on a sloping shore, then you can make the targets of fixed pre-measured height with a nice solid base, and simply place them with the base on the waterline. Would need some adjustment for slope though
 
Wallace's correspondence from the time is still relevant now.

http://people.wku.edu/charles.smith/wallace/S179AA.htm
"I have had much difficulty in getting a telescope of sufficient power. I have one that may do (3 inches object glass and 4 feet long), but should prefer one of 4 inches object glass and five feet long; and these are only made for astronomical purposes. I can only find onein London, and that will not be let out on hire or trial, and only sold (without stand) for £42. I believe, however, I can get the loan of one from Brighton if mine turns out insufficient, and if I cannot get one at Lynn. I have had one telescope on trial already, and found it not good enough.
Content from External Source
Six miles is a very long way. However I think a P900 or similar would be sufficient. Here is a zoom test of 5km (3.1 miles)

Source: https://www.youtube.com/watch?v=VQPZuBlCvvg

20160913-075328-c554f.jpg
20160913-075450-0390i.jpg

So that's the type of angular resolution and distortion you would get at the midpoint. Sufficient to resolve railings that are likely only an inch thick or less.

The far point, at 6 miles, would be half the size, and with additional atmospheric distortion. Wallace used "a large sheet of white calico, six feet long and three feet deep, with a thick black band along the centre". I'd suggest the same, and based on the above a 3" thick black band. Specifically I suggest using two strips of Gorilla Tape, which besides being black is also very unlikely to fall off the board (and black duct tape could do, Gorilla Tape is more heavy duty). The board itself could be a piece of plywood painted white, or some poster board (might be too fragile), or even an actual dry-erase whiteboard.

I also suggest giving the board a distinctive border with tape, and maybe some colored squares to make it easier to find, like:
20160913-081830-64335.jpg
 
Last edited:
A series of three bridges such as in the original seems ideal, as you can use a simple plumb measure to get the height.

The original Bedford Level is still there, and has been used to kind of recreate the useless "man in a boat" version of the experiment. But it's not clear if it's still suitable for the Wallace experiment. These seem to be the bridges:
20160912-141904-gg5f7.jpg (See attached Google Earth File)

However the middle bridge is a railway bridge and does not seem to have access.

There are footpaths/bridleways along both sides of the Old Bedford River all the way along: http://streetmap.co.uk/map.srf?x=55...ty/Town/Village]&searchp=ids.srf&mapp=map.srf

Street View suggests that overhanging trees might be a problem, though.

upload_2016-9-13_16-48-45.png


There's also a dead straight 6 mile stretch in the other direction from the Welney bridge to the next bridge:

upload_2016-9-13_16-47-26.png
 
Grand Traverse Bay in Michigan would be a good contender.

There's a peninsula in the middle, and it's at about the halfway point.

For some reason, my computer doesn't want to screenshot it.

Anyway, it's about ten miles from the peninsula near sutton's bay to the mainland, and where the Old Mission Light is is roughly halfway. You could put the targets in water, maybe have them mounted on a 5-gallon bucket filled with concrete, put a 'water level' mark on them wade them out that far, and presto. Easy way to replicate the Wallace experiment, and if you made them ten or twelve feet high, your'd potentially minimize the effects of refaction near the surface.

Now, if you had a boat, you could make floating targets (a tripod supported on three inner tubes) and use your GPS to space them however many miles apart you wanted for the experiment, and you could really do that nearly anywhere on the Great Lakes.
 
There are quite a few dry lake beds in California and Nevada. How do you tell the surface is exactly level? When the winter rains come along, it's common for a lake bed to be covered with a thin layer of water only inches deep. You can still drive over the lake bed, but on a still day the surface of the water should be a very good gravity equipotential surface.

As for the Bedford Level, I've previously asked people on another forum if there were any UK members interested. I got some casual interest. I'll ask again, more seriously.
 
Am keeping an eye on this.. I am based about 1hr drive from the original Canal. Not sure how involved I can be directly though as lots of family commitments :eek:(
 


So that's the type of angular resolution and distortion you would get at the midpoint. Sufficient to resolve railings that are likely only an inch thick or less.

I just got the same camera, a Nikon P900, mostly for videoing planes, but it's also ideal for earth curvature experiments with a 2000mm equivalent zoom. It's also much more practical that a DSLR with a long lens, as it's much lighter and smaller. Here's an "actual pixels" handheld shot of a gas station awning about 5km (3 miles) away. It's a hot day (around 85°F)

20160928-100140-fnjax.jpg

Here's the same type of gas station
20160927-165952-cezoi.jpg

So you can see you get reasonable contrast with the approx 1 foot blue stripe on a 2 foot white background. The red and blue Arco logo is also reasonably visible.

20160928-100324-w4ycd.jpg
Remember though that the Wallace experiment was twice as far as this, so this represents a bit past the midpoint.
 
Last edited:
It was a bit hot for the time of year here. I'll try taking photos of the same scene and recording weather conditions, to see what the effect is. There's some varied weather coming.

20160927-173346-8hmpm.jpg
 
I'll try taking photos of the same scene and recording weather conditions, to see what the effect is.
You also can try taking short HD videos of the scene. Although there will be fewer pixels per frame, by going through the video frame-to-frame, it is possible to find a frame with a sharper image than in a single still photo. I found this trick helpful when trying to read the tail number of a distant aircraft.
 
You also can try taking short HD videos of the scene. Although there will be fewer pixels per frame, by going through the video frame-to-frame, it is possible to find a frame with a sharper image than in a single still photo. I found this trick helpful when trying to read the tail number of a distant aircraft.

I took a short video of the freeway about 300 feet from the gas station. Like you say there are occasional moments of clarity from the camera movement and the heat haze (turbulent atmospheric refraction)


Source: https://www.youtube.com/watch?v=ihn9kANb5Qo


A tripod will help :) But I think it's clear that a midpoint target need only be around 1 foot in diameter, providing it is of sufficiently high contrast (like a red disc).
 
Although, shot this early morning so I could see the gas station lights. Did not come out incredibly well handheld

20160928-100535-6tgfj.jpg

Still, the strip of blue lights are quite distinguishable.
 
Last edited:
A tripod will help :) But I think it's clear that a midpoint target need only be around 1 foot in diameter, providing it is of sufficiently high contrast (like a red disc).
Actually, I was thinking about the noise reduction by taking multiple images of the same scene (with help of a tripod mounted camera) and averaging them. Taking a video probably is the easiest way of doing it. There may be burst and continuous shooting options, but they require holding the shutter button that may cause the camera drift.
 
Actually, I was thinking about the noise reduction by taking multiple images of the same scene (with help of a tripod mounted camera) and averaging them. Taking a video probably is the easiest way of doing it. There may be burst and continuous shooting options, but they require holding the shutter button that may cause the camera drift.

That's in interesting idea, I know you can increase effective resolution, as explained here:
http://petapixel.com/2015/02/21/a-practical-guide-to-creating-superresolution-photos-with-photoshop/
By taking a burst of numerous consecutive photographs by hand and cleverly combining them in post processing, we can noticeably improve the resolution capability of any camera. It’s a simplified geometrical reconstruction technique using the concept of sub-pixel image localization.
Content from External Source
However there's some suggestion it might not work well with atmospheric distortion.
http://www.cambridgeincolour.com/forums/thread44497.htm#post492514
The technique depends on the noise not being common to each frame. Fine in principle, but actual experience was that the final stack was invariably softer than any of the original frames. I came to the conclusion this was because atmospheric distortion actually changed the absolute dimensions of the moon in each frame by a few (sometimes more than a few), pixels.
As a result fine detail, visible despite noise in the individual frames, disappeared. (presumably because it starts to resemble real noise.)

At the end of the day it was better to process individual frames for noise and properly apply selective sharpening using luminosity masks to target the fine detail.
Content from External Source
I'll tripod the camera and give it a go.
 
So I'm trying to figure out best time of day. Here's a shot at first sunlight:
20160928-100803-y3v2d.jpg
Still rather a lot of atmospheric distortion. It was about 70°F/21°C
 
Last edited:
Overcast days seem best. Pity about where you live..
Yes it's terrible here. Warm and sunny all the time, and worst of all, no contrails for the last two months!!

Overcast would be ideal as there would be basically no heating of the ground, greatly reducing convection turbulence. Maybe later in the week.
 
Well, a tripod and more sunlight made an astonishing difference, 9:20, around 80°
20160928-100941-8pw56.jpg

That's on a tripod with manual shutter release (i.e. I pressed the button) and image stabilization on

Compare with yesterdays afternoon handheld


Note the difference in the red and blue signs. Barely visible in the older shot.
 
Last edited:
I suspect a large part of the reason why the above image is so good is the closeness of air temperature and ground temperature.

Outside right now, measure the shaded air temp at 76 and the ground temp at 72 (using a Thermopen meat thermometer both in soil and an unheated spa). The sun has not yet heated the ground sufficiently to cause it to warm the air above it (and create convection).
 
Well, a tripod and more sunlight made an astonishing difference, 9:20, around 80°
20160928-100941-8pw56.jpg

That's on a tripod with manual shutter release (i.e. I pressed the button) and image stabilization on

I can't quite tell from the crop, are those wires at the gas station (or at least close by)? That's pretty incredible to be able to pick out individual utility cables from 5km away.
 
Back
Top