The "Whatever Theory" Identifying The World...

[Mike Smith Cosmos]

Whatever theory

 

Just an interesting observation which , May or may not , have any use, in your discussions .

 

We identify a car we own by a registration plate , say WYO 774 . Which allows for reasonably a large amount of individual numbers , so that I do not have the same number as a " Robber " and end up in prison by accident !

 

But your Colour coding principle sounds like it can provide and " Humongous " amount of identification colour 'tabs' . If nature has chosen colour as being a distinguishing feature for ' very good reason ' it will be very interesting ,how this pans out.

 

Light seems fundamental to the operation of the universe, and as Isaac Newton showed with his prism experiment :- light is a combination of a rainbow full of different colours, which current spectroscopy has further investigated into a very ,very large science .

 

Mike

[tar]

Whatever Theory,

 

The amount of numbers we have is considerably reduced, if we go by tint and not shade. If we go by "ratios", R220 G40 B148 has the same hue as R200 G20 B128 or R180 G0 B108. So the conventional, one number representation of R/B 1.7, represents a whole bunch of shades of the same hue, thus putting all those many numbers under one identifiable shade and reducing our 16 million "different" RGB values to a much smaller stable of hues.

 

https://en.wikipedia.org/wiki/Pantone

 

Is an example. The "spot colors" that Pantone makes out of 14 base pigments total 1,114. Many of these, especially the "creamy" ones with white pigment in them, can not be produced with Cyan Magenta and Yellow. So the hues that I could "come up with" by using the undercolor reduction principle are a very limited number of hues, that would not allow for a different number for each species.

 

Better however to work with 100 or 500 or 1000 identifiable hues than to just say the flower is purple, or the butterfly is black and white. And even if you only had 100 different hues, 2 color areas would still give you 10,000 combinations. Three color areas would give you 1,000,000 and 4 color areas would give you 100,000,000, so even if you loose the 16,000,000 colors in favor of 100 hues you might still be able to have enough unique 4 or 5 color codes, for each specie.

 

You have to nail down the scheme though, and come up with the rules of how to come up with a single number for a color patch that has all sorts of different RGB triplets in it.

It has to be such that if you analyze a patch, and I analyze a patch, we come up with the same number Mike would come up with, and even more importantly, the same number a smart phone would come up with, if it executed the instructions we gave it, to come up with a number for the color of the patch.

 

Regards, TAR

[Mike Smith Cosmos]

.

 

Here , we in our cognition , are able to say 'yes '

 

This is a sailing boat , yes they can have a hull that is soft green ( though unlikely ) sail white , fairly normal sea blue yes but a bit odd reddish . Sky odd should be blue yet it could be sunset . Ok.

 

 

No way this boat is all violet , violet reflection , no this is all wrong , sail indigo not really unless It was a London barge . But yes the sea and the sky look o.k. Well I do not know what to make of this.

 

So we use colour in our perception and identification of things . Or not as the case may be.

 

 

Mike

Edited July 6, 2015 by Mike Smith Cosmos

[Arete]

Species delimitation is a field in which I have some expertise, so I think I can offer a few insights.

 

Color can sometimes be a useful delimiter of species. However, on its own, is generally considered inadequate. There are a few major reasons why this is the case.

 

1) Color often shows high levels of intraspecific variation - i.e. individuals of the same species can often exhibit different coloration. E.g. these are all color polymorhpisms of the same species:

 

 

2) Cryptic species generally do not show distinct coloration, despite being distinct species. E.g., these are all distinct bumblebee species:

 

 

3) Singular data forms of any kind are generally considered inadequate to delimit taxa. Generally, a multivariate approach is considered necessary to do so these days. Multiple morphometric, meristic, genetic and ecological datasets should be collected, and analysed using a multivariate approach, giving axes of differentiation between populations and species.

 

This is not to say that categorizing organisms by color is always a futile exercise, as sometimes it can yield interesting insights:

 

http://onlinelibrary.wiley.com/doi/10.1111/j.0014-3820.2004.tb00462.x/abstract

http://onlinelibrary.wiley.com/doi/10.1111/j.1558-5646.2008.00507.x/full

http://www.sciencemag.org/content/313/5789/963.short

 

But we already know its unreliable as a singular datum to identify and catalog biological entities.

[whatever theory]

Hello Mike, thank you for your comments.

 

 

But your Colour coding principle sounds like it can provide and " Humongous " amount of identification colour 'tabs' . If nature has chosen colour as being a distinguishing feature for ' very good reason ' it will be very interesting ,how this pans out.

Nature has provided a unique color range for each species so that it can be identified, similar to your license plate, from a distance.

Color is consistent threw out nature and no where is it not found. This should show that color may be natures number one means of identification.

 

 

 

Light seems fundamental to the operation of the universe, and as Isaac Newton showed with his prism experiment :- light is a combination of a rainbow full of different colours, which current spectroscopy has further investigated into a very ,very large science .

These next questions are just food for thought, maybe they are for another thread or time, but since we are on the subject of color I will ask them, but do not expect any replies.

1. What color is light?

2. What color does a star mostly resemble?

3. What color is one of our most abundant minerals, salt?

4. What color would be first to have evolved, in life?

5. What is the one color that most all living creatures share?

6. What is the second most common color most all living creatures share?

7. What could the answers of these question have to do with how life evolves from the very beginning?

 

Maybe some day down the road I will share my thoughts on these questions. But the answers are now already available in our collective consciousness, so as Tar pointed out, you can all already tap into the answers, if you want.

 

Very nice artwork Mike and as always your comments are enlightening and inspiring.

 

Thanks Tar for your comments

 

Better however to work with 100 or 500 or 1000 identifiable hues than to just say the flower is purple, or the butterfly is black and white. And even if you only had 100 different hues, 2 color areas would still give you 10,000 combinations. Three color areas would give you 1,000,000 and 4 color areas would give you 100,000,000, so even if you loose the 16,000,000 colors in favor of 100 hues you might still be able to have enough unique 4 or 5 color codes, for each specie.

You have to nail down the scheme though, and come up with the rules of how to come up with a single number for a color patch that has all sorts of different RGB triplets in it.

It has to be such that if you analyze a patch, and I analyze a patch, we come up with the same number Mike would come up with, and even more importantly, the same number a smart phone would come up with, if it executed the instructions we gave it, to come up with a number for the color of the patch.

I see what you mean. I am going to try to think of a way to incorporate all of this into one technique.

Maybe I will have some kind of solution tomorrow.

Thanks again

 

Thanks Arete for your comments and for participating in the discussion. Your expertise on these matters is very much appreciated.

These bees are a perfect example of how this technique can be beneficial.

Here you have 8 bees, of different species, that all appear, to the naked eye, to be the same color.

Not only do they appear the same color as each other, but they all have most of the same characteristics other then color such has hair, shape, size, etc, to each other.

So in this case identification of the species is going to be very difficult.

If you notice the orange stripe closest to the bees rear end has the most amount of color variation in each bee.

You can notice that this stripe is very different from one bee to the next. Some have almost a white and some are very dark. Some are very yellowish and some are more orangish.

I have found the lowest Red RGB value for each bee which is the top set of numbers above each bee.

Then I have found the highest Red RGD value for each bee, in this same rear end orange stripe, which is the bottom set of numbers above each bee.

If you notice there is a clear difference in the highest, lowest and the ratios in between each bee.

The next test I will perform I will try to compare bees of the same species to each other.

The picture that you posted of the grasshoppers I did not want to touch, at all, because if you will notice the back ground in each picture is different colors. So I do not think the conditions are the same in all the photos.

But I do see a pattern emerging from these grasshoppers between each other.

The 2 on the bottom appear to have mostly the same shades of green and the upper right grasshopper appears to have these same shades of green on the bottom of his legs and this grasshoppers yellowish color appears to be the same shades as the one in the upper left hand picture.

The 2 most important factors in determining species that I can think of are:

1. Geographic location and this can mean a lot of things.

If say a wolf lived in Miami this same species of wolf that lived in Alaska would probably appear very different.

But if a wolf lived in Montana it may share the same color as the one that lived in Alaska because the conditions are so similar.

South Sea pearls that are grown in Australia, Philippines, and Indonesia usually have the same color ranges as each other but they do not resemble these same species from say Mexico.

If there was a certain area that had a crowded population of grasshoppers, all of these crowded grasshoppers may all share color ranges with each other, but may not share the same color ranges as another colony of these same grasshopper a mile away that was not crowded.

2. Seasons may have some thing to do with the difference in species color and sometimes they may not.

For instance the South Sea pearls I was just talking about. They appear to be the same color no matter what time of year they are harvested, because there are no real seasons, this close to the equator, and all of the other conditions are almost identical in these 3 locations.

But in Alaska and Montana it will depend on what time of year you see the wolf, that will determine their color range.

 

Here are the first 12 pictures I could find on Google images of pairs of grasshoppers from the same species, same geographic location, and the same time of year.

You will notice that they all share the same color ranges as there same species pair.

 

 

 

Thanks again

P.S. I hope these Grasshopper pictures are not too offensive for this forum.

These pictures are for you Mondie. "colorful sexual displays" OOOOO Yaaaaa....

Edited July 7, 2015 by whatever theory

[tar]

Whatever Theory,

 

I gave you a plus on the bees.

 

I had honed in on the same stripe, as a potential differentiator in terms of color, and am glad you went after it.

 

But, we still have to work on the "ratios". There is something important there, but we have to call it right, and call it consistently to see what it is that is there.

 

For instance in your low red triplet of R95, G48, B12 you figure an r/g ratio of -47. That is true that there is a difference of 47 between R and G but 95/48 is not -47, it is 1.97916, so if you are going to state an R/G value, related to that triplet, you might state that it is 2.0 (if we round to one significant digit to the right of the decimal point.)

 

The high red triplet you found for the first bee was R204, G159, B119. You stated the R/G ratio was -45. But 204/159 is (rounded) 1.3.

 

As you can see, the R/G ratio changes significantly between the numbers in the low red triplet and the high red triplet...but the relative amounts of the red and green do stay the same, in terms of the undercolor reduction we were talking about, or to put it another way, the low triplet might be close to being a lighter shade of the same hue that is represented in the high triplet.

 

To see the relationship between Red and Green, lets take out some white light, from the triplet. We are starting with R95, G48, B12, so lets take out R12,G12,B12, leaving us with R83,G36,B0.

Here the R/G value is 83/36 or 2.3.

 

The high triplet was R204, G159, B119, so lets take out some light...R119, G119, B119, leaving us with R85, G40, B0 which is an R/G ratio of 85/40 or 2.1

 

If you took 36 samples of the color in that band of the first bee and performed the same R/G ratio calculation and averaged the results, you might come out with an R/G 2.2 hue to describe that band. One thing you might do, short of the 36 is take 3 samples and arrive at a number. then take a different three (random)samples and arrive at a number, and then average the two. Then take a different three and come up with a number, and average the three together. Does not matter how, but see if the R/G ratio, tends toward 2.2 or 2.3 or 2.1 or 1.9 or whatever, but it should tend toward some number. Since we have computers and the RGB value of each pixel is already recorded, once an area is defined, you could add up all the red values, divide by the number of pixels in the area and get a Red value, same for Green, same for Blue. Then which ever value is lowest, take the light out Rx, Gx, Bx that would zero out one of the colors, and then take the larger of the remaining numbers, divide it by the smaller of the remaining numbers and you would have an R/G value or a G/R value or an R/B, B/R, or a G/B, B/G value.

 

Two samples obviously are not likely to show a single number, as in the first trial of the system we got R/G 2.3 and R/G 2.1, but using the system, if you were to take ALL of the pixels in the area, you would come up with only one number. The question would be, can you come up with this ACTUAL average ratio, by taking the high red triplet and the low red triplet, finding the ratio between the two strongest colors(after taking grey out, to leave just the two) in each triplet, average them together and get the ACTUAL average ratio...or do you need 3 or 5 or 10 or 36, or 100 or 1000 random samples to get the ACTUAL average pixel in the target area.

 

Let's try the WT-TAR scheme on the second bee. The first bee yielded R/G 2.2. The second bee has the low red at R81, G50, B30. Taking out white light we get R51, G20, B0, or an R/G of 2.6. Taking the high triplet for the second bee of R161, G94, B10 and removing R10, G10, B10 we get R151, G84, B0 or an R/G of 151/84 or 1.8...Not enough info to call the one number (bee one's band number) identifiably different than the other (bee two's band number). In fact 2.6 and 1.8 average out...to 2.2.

 

Maybe we just identified a certain amino acid substitution variant of the Mclr molecule present in both bees, but we didn't give that band on bee 1 a different WT-TAR number than the band on bee 2. To find the ACTUAL average hue on the band we need more sample triplets. Enough to be satisfied that averaging ALL the pixels would give us the same number we arrive at, by averaging the B/G of the low red triplet and the B/G of the high red triplet.

 

Regards, TAR

although I do get R/G 1.7 for bee three, (1.6 and 1.8) so maybe the 6 species do each have a different ACTUAL R/G number for that band. We need more numbers for one and two, to be satisfied we have the ACTUAL average hue of the band for each, and to determine if those numbers alone could differentiate the species.

Edited July 7, 2015 by tar

[tar]

...Later

 

I recently obtained a Samsung S4 smart phone and joined the 21st century.

Yesterday I installed a free app from Play Store called pipette that lets you pick a picture in your gallery, zoom in on a section, lock it, and then move your finger around, and a little circle rides above your finger, and the RGB value of the pixel in the center of the circle appears below in decimal and hex.

 

I saved Whatever's version of Arete's 8 species of bees and performed the TAR modified version of Whatever's color code method, to obtain a different R/G value for bee one's band and bee two's band.

 

I zoomed in, so the entire single bee was in the picture, locked it, and entered 6 values into an Excel spread sheet, that totaled and averaged each color, subtracted the Blue average from each average, and divided red by green.

 

For bee number one I ran the six point trial 4 times and got R/G= 2.1, R/G=2.0, R/G=2.0 and R/G=2.1.

For bee number two I ran the six point trial 4 times and got R/G=1.8, R/G=1.9. R/G=1.8 and R/G=1.8

 

This is a technique that requires some more trials to verify that it results in the same number every time the same area is checked, but it looks to me, at this point, that bee one's average hue on that band is going to be close to R/G 2.05 and bee two's avg hue on that band is going to be close to 1.83.

 

This amounts to a 24 point check, but it looks like it might provide an identifying number.

 

I did not retain the numbers for bee one's trial but the last thee sets of 6 for bee two's lower light band, yielded these three average hues. R143,G86,B25 R141 G91,B27, and R142,G88,B21.

Those numbers should yield R/G 1.9, 1.8 and 1.8 if I figured and recorded right. For instance, the first would be after subtracting R25, G25, B25... R/G 118/61= 1.934 rounded to 1.9.

 

Try it.

 

Regards, TAR

[MonDie]

This is my main motivation behind all of my work "colorful sexual displays" OOOOoooo Yaaaa!!!...

 

I guess it's my theory then.

 

---

 

Species delimitation is a field in which I have some expertise, so I think I can offer a few insights.

 

Color can sometimes be a useful delimiter of species. However, on its own, is generally considered inadequate. There are a few major reasons why this is the case.

 

1) Color often shows high levels of intraspecific variation - i.e. individuals of the same species can often exhibit different coloration. E.g. these are all color polymorhpisms of the same species:

 

 

2) Cryptic species generally do not show distinct coloration, despite being distinct species. E.g., these are all distinct bumblebee species:

 

 

3) Singular data forms of any kind are generally considered inadequate to delimit taxa. Generally, a multivariate approach is considered necessary to do so these days. Multiple morphometric, meristic, genetic and ecological datasets should be collected, and analysed using a multivariate approach, giving axes of differentiation between populations and species.

 

This is not to say that categorizing organisms by color is always a futile exercise, as sometimes it can yield interesting insights:

 

http://onlinelibrary.wiley.com/doi/10.1111/j.0014-3820.2004.tb00462.x/abstract

http://onlinelibrary.wiley.com/doi/10.1111/j.1558-5646.2008.00507.x/full

http://www.sciencemag.org/content/313/5789/963.short

 

But we already know its unreliable as a singular datum to identify and catalog biological entities.

 

My photoreceptive speciation theory predicts that bumblebees cannot detect differences in hue, and likewise for the grasshoppers unless the grasshopper species is undergoing speciation. If a species can distinguish hues, color distinctions will develop between species (unlike the bumblees). Furthermore, color schemes will be conserved within species (unlike the grasshoppers), and the appearance of different color schemes will indicate that speciation is occurring. The bumblebee's cryptic coloration did not override, but it was maintained because color had no other function.

Edited July 7, 2015 by MonDie

[tar]

Mondie,

 

Then why would flowers be different colors? They don't even have eyes to detect a hue at all.

 

It seems mutations that change the Mclr molecule, or whatever wavelength absorbing molecule a species has, can be selected for or against not based solely on a mate liking the display, but based on the effect the color change might have on OTHER species, that might pollenate a flower more, or keep a predator from eating you, or any other survival tendency the mutation might provide.

 

Regards, TAR

maybe that bumble bee coloration is very effective at having clover grazing animals, NOT step on you, AND blends you in with the yellow flowers you're feeding on, against the watchful eyes of hungry birds

[whatever theory]

Hi Tar

I really like your new formula. Thank you very much for taking so much time to figure it all out. From now I will refer to this as the "TAR Formula."

I was thinking that maybe a good way to find the numbers is to do as I have done previously and take the highest Red and the Lowest red value.

Then Find out exactly what color should be in the exact middle. Say the high value is 160 and the low value was 100 then the exact middle would be 130. Then if you wanted to take more reading you could find the midway point in between 100 and 130 then 130 and 160.

So now you would have values at: 100, 115, 130, 145, 160.

Then you could try to match these 3 midway numbers to the other subject or you could perform the same color range test on the other subject.

Please let me know what you think.

I downloaded the Pipette software, but I can not figure out how to load a picture into it. It is just a black screen. What do I do next?

That would be great if there were an easier way of doing this....

Thanks again

 

After thinking about the plant identification Apps I have answered one of my own questions about it.

 

 

If you did not take your picture at the exact same distance from the subject every time or if your zoom was different in different pictures, then your sizes would be all over the place. So this may not be a good deciding factor.

 

By using a piece of grid paper they are not only able to use this white grid paper as a white color control, but they may be using the grid to determine the species size by comparing the species to the size of the grid.

Since I do not find size and shape being of much importance to a future App, using primarily color, this would not be able to add anything in this fashion.

But if in the future we were trying to compare pictures that were taken in normal conditions, instead of just using a white or grey card in the picture, we could use a white/grey card that also had a grid on it.

This grid would help you determine what distance you would want to take your picture from the subject, based on the grid.

This way you would always be able to take your pictures from the same distance, as previously.

 

I thought of a good example of when species will not share the same colors even if they are all in the same exact geographic locations.

Lets say, that Mike owned an Alligator Park in the Florida everglades. "Mike's Alligator Sanctuary"

Tar is the Chief Alligator Wrestler and he takes very good care of his gators and he feeds them a very special diet to make sure that they get all the nutrients that they need, but it is still not the same diet that the other Alligators, of the same species eat, outside of the park.

Tar keeps there pools nice and clean and tries to replicate the waters outside the park, but this water is still not natural.

Mondie is in charge of the sexual breeding program and is trying to breed very rare blue spots into his Alligators because the females are more attracted to these blue spot inside the park, where as outside the park females are trying to get rid of these blue spots by not choosing these freaks to breed with. That is why it is rare in the wild to see these blue spots.

Studiot is the gardener and is trying to improve the red color of native plants, brought into the park, with special fertilizers.

So now while all of these Crowded Alligators and flowers are in the same geographic location as the Non-Crowded Alligators and flowers just outside the park,they will have different colors then them.

So instead of comparing them to the others outside the park you would only compare them to others inside the park.

So in this case your location would be "Mike's Alligator Sanctuary" and not the "Florida Everglades."

 

I am going to add a few question to my "Food For Thought" List, maybe now you will start to see a pattern emerge.

 

1. What color is light?

2. What color does a star mostly resemble?

3. What color is one of our most abundant minerals, salt?

4. What color would be first to have evolved, in life?

5. What is the one color that most all living creatures share?

6. What color is associated with bones and teeth and the outer part of an eyeball?

7. What color is associated with Brains?

8. What color is Space?

9. What is the second most common color most all living creatures share?

10. What color is the inner part of an eyeball?

11. What color is associated with a Tongue, Heart, Blood, Dirt, Lava, Fire?

12. What color is associated with Plants and most Reptiles?

13. What color is associated with the Sky and Water?

14. What could the answers of these questions have to do with how life evolves, from the very beginning?

Edited July 8, 2015 by whatever theory

[whatever theory]

[DrP]

1 - white

2 - depends on the star

3 - depends on the mineral.. oh - salt - clear (single crystal). (although can appear white depending on the size of the crystals and number).

4 - ? No-one was there to testify.

5 - Doesn't make a lot of sense to me... bone?

6 - Depends on the creature - but bone white.

7 - 'associated' - grey. Actual - don't know.

8 - non question really - depends what it is filled with.

9 - hmm - don't know why I don't like this question - seems objective and depends on definitions etc..

10- I think it depends on the animal - I am not sure

11 - is there any relevance to this? reds, browns and yellows.

12 - green

13 - The sky is blue due to scattering and water is blue due to free electrons

14 - some of them have nothing to do with it.

Edited July 8, 2015 by DrP

[Strange]

1. What color is light?

It depends on the source.

2. What color does a star mostly resemble?

It depends on the type and age of the star.

3. What color is one of our most abundant minerals, salt?

No colour

4. What color would be first to have evolved, in life?

Who knows.

5. What is the one color that most all living creatures share?

Pink?

6. What color is associated with bones and teeth and the outer part of an eyeball?

Greyish pink

7. What color is associated with Brains?

Greyish pink

8. What color is Space?

No colour

9. What is the second most common color most all living creatures share?

Blue?

10. What color is the inner part of an eyeball?

Transparent/no colour

11. What color is associated with a Tongue, Heart, Blood, Dirt, Lava, Fire?

Red and black

12. What color is associated with Plants and most Reptiles?

Green and grey respectively.

13. What color is associated with the Sky and Water?

Blue and none, respectively.

14. What could the answers of these questions have to do with how life evolves, from the very beginning?

Very little. Although, plants on a planet round a red dwarf, probably wouldn't be green as there wouldn't be much short-wavelength light. They would probably be nearly black.

Edited July 8, 2015 by Strange

[DrP]

Water IS blue Strange.. ;-) Although only very faintly. :-D It's due to the free electrons in it... unless this is yet something else we were taught at school or uni that turns out to be false or out dated. There is nothing on wiki about the blue in water being from free electrons, but that happened a lot. ;-)

[Strange]

Water IS blue Strange.. ;-) Although only very faintly.

 

I agree. When I was on a glacier surrounded by ice, it was stunning how blue everything was.

 

But I'm not sure blue is "associated" with water - think of adverts for pure,clear spring water and mountain streams. (Or grey English rain!)

[tar]

Whatever Theory,

 

Finding the high and low red, gives you too wide a range, and to pick a number directly in the middle might not be the best way to represent the color. Especially because you don't know if the other colors are going to work in the same manner. That is why I chose to avg the numbers individually, but use numbers from the same randomly selected triplets, to do it. No selection bias. And suppose you "missed" a pixel that had a higher or lower red than you selected. Anyway, whatever system you use should come up with some number, some one number that any person or program would arrive at, by looking at the same patch. Otherwise, you have nothing to put in the database as the number for the color of the band on that species of bumblebee.

 

On the pipette program, I forgot to mention a step. I downloaded your picture of the bees to my desktop and emailed it as an attachment from my gmail account to my yahoo account. Then I opened the email on my smartphone and downloaded the attachment into my gallery. The pipette program had a plus button that you push which allows you to bring a picture from your gallery into the pipette program. Then you position the part of the picture you want to inspect, zooming in, if you like, and hit a little "lock" button up on the upper right of the screen. Then you will have the little circle above where you touch that displays the RGB value of the thusly targeted pixel.

 

Regards, TAR

[MonDie]

New criteria for the photoreceptive theory:

- can detect differences in hue

- reproduce sexually and avoid inbreeding

Bees actually see a broader range of color than we, but bees don't actually search for mates. Per the link below, the drones fly out of the nest chasing the evacuating queen, and it's here that they compete for her. http://insects.about.com/od/antsbeeswasps/qt/Honey-Bee-Mating.htm They inbreed, mating within the colony. If this mating strategy rules out hybridization, then we have an exception.

Mondie,

 

Then why would flowers be different colors? They don't even have eyes to detect a hue at all.

 

It seems mutations that change the Mclr molecule, or whatever wavelength absorbing molecule a species has, can be selected for or against not based solely on a mate liking the display, but based on the effect the color change might have on OTHER species, that might pollenate a flower more, or keep a predator from eating you, or any other survival tendency the mutation might provide.

 

Yes, the model works for flowers if you think of the bee as an extension of the flower involved in fertilization. Apparently bees exhibit flower constancy, meaning they consistently choose the same flower species. This means that a bee-pollinated flower might recieve more pollen by distinguishing itself from other closely related species. Thus the cycle of color exaggeration begins, that is, at least, according to this highly speculative theory I just made up.

[Arete]

You may find this software interesting: http://www.naturepatternmatch.org/Using color to infer evolutionary processes is not a new concept.

 

However, I think you're misleading yourself by pitching it as a singular method to delimit species/populations. We already know that it is unreliable for that purpose.

 

These bees are a perfect example of how this technique can be beneficial.

 

I tend to disagree. Any six individuals are extremely unlikely to be exactly the same color, so being able to delimit individuals without being able to segregate individuals into predetermined categories (i.e. taxonomic units) is of very limited utility.

 

To be more succinct, the issue with trying to delimit taxonomic groups of organisms solely on the basis of coloration is that variation within species tends to overlap with variation between species. There's a natural distribution of coloration in any group of organisms, and in related species, they often tend to overlap. To display this graphically, say we have species A in blue, and species B in red. Species A tends to be color 1, Species B tends to be color 2, but there's a proportion of the distribution of both species (shaded in purple) where the observed color distribution overlaps. Individuals falling into this region of each species distribution will be mis-classified using a singular approach based solely on color, no matter how sophisticated the approach is. Ergo, using a singular dataset (in this case, coloration) to try and classify such a group of organisms is fundamentally flawed, and will never function well.

 

 

The most accepted approach, as I previously stated, is to use a multifactorial approach taking into account various data forms, such as morphometrics, genetics, and the physiological limits of a given species to delimit taxa. Coloration can be a component of a multifactorial approach where applicable, but it's unlikely to be taken seriously if pitched as a sole delimitation practice.

 

To conceptually frame your method in current evolutionary thinking, I'd strongly suggest a read of this paper by Kevin De Queiroz on species concepts http://www.ncbi.nlm.nih.gov/pubmed/18027281 and, hopefully it's not too presumptuous to self cite here, but here's an example of how multifactorial species delimitation works in practice I published a couple of years ago. https://sistromlab.files.wordpress.com/2015/03/sistrometalmolphyecol.pdf

Edited July 8, 2015 by Arete

[tar]

Mondie,

 

So, might it make sense that a particular yellow, that had a particular WT-TAR hue might exist in both the flower and the associated bee?

 

Regards, TAR

 

Once we come up with a standard method to arrive at a repeatable number. Whatever Theory or someone else will have to see what those pictures of yellow bees on yellow flowers tell us.

[Strange]

 

So, might it make sense that a particular yellow, that had a particular WT-TAR hue might exist in both the flower and the associated bee?

 

This is complicated by the existence of mimicry. Other plants may mimic the preferred flower in order to attract pollinators, on in the case of carnivorous plants, prey. They may also mimic the insect (*). Other animals may mimic the insect either because they are predators who don't want to be spotted or because they use it to scare off potential predators (e.g. hoverflies, which look almost identical to various species of bees and wasps).

 

(*) https://xkcd.com/1259/

(This is not completely accurate, but it is a nice story...)

Edited July 8, 2015 by Strange

[tar]

Strange,

 

Talk about unrequited love!

 

Regards TAR

[tar]

Arete,

 

I suggested to WT in a PM that we try to falsify his theory by each doing a 36 random point check of that band in each of your 8 bees. I asked that he tell me if he is game, before I do the work to come up with the data points and the calculations. I have not heard from him yet, but I started the trials, and here is the first trial of 6 random points, from bee 3 and the same from bee 8, along with WT's "high red" and "low red" triplets. Notice my random picks found a higher red in the one bee than he found, and a lower red in the other, but he was staying away from dark areas and glare areas, and I just picked on a loose 2x3 grid basis. I had picked a single triplet from each bee, just to get a feel for which bees might be the closest My single pixel quick check yielded 1.7, 1.7, 1.6 , 2.2, 1.8, 1.5, 1.9 and 1.6 respectively. I picked 3 and 8 because they looked like they would be likely to yield close to the same number, when WT did his 36 point check, and I did my 36 point check.

Interestingly 6 points suggested there may be an identifiable color difference in the bands between 3 and 8. I am interested in WT taking up the challenge, I will continue to post random triplets and R/G numbers as planned, and I encourage others to run a similar 6 point, 12 point, 24 point or 36 point check on at least bee 3 and bee 8. If we all can agree on a single number for the band in bee 3, and a single number for the band in bee 8, it won't prove the theory, might falsify it, but might indicate that further testing is in order.

 

1^st trial of Six point random check

 

Bee 3 band R184 G119 B18

R165 G101 R/G=1.63

 

TAR Data for band 3 174|112|3(R/G= 1.6)| , 169|103|17(R/G=1.8)|, 178|118|26(R/G=1.7)|, 209|141|32(R/G=1.6)|, 179|117|0(R/G=1.5)|, 192|124|31(R/G=1.7)| average TAR points 1.64

 

WT Data for band 3 104|73|20 (R/G=1.6), 185|124|49 (R/G=1.8) average WT points R/G=1.7

------------------------------------------------------------------------------------------------------------

Bee 8 band R153G99 B32

R121 G67 R/G=1.81

 

TAR Data for band 8 162|111|20(R/G=1.6), 165|103|16(R/G=1.7), 153|100|46(R/G=2.0), 126|78|20(R/G=1.8), 182|112|52(R/G=2.2), 128|91|38(R/G=1.7)

 

WT Data for band 8 141|91|29 (R/G=1.8), 213|159|97 (R/G=1.9). WT together R/G=1.8

Edited July 9, 2015 by tar

[Arete]

Arete,

 

I suggested to WT in a PM that we try to falsify his theory by each doing a 36 random point check of that band in each of your 8 bees.

 

A photo of single individuals does not take into account intraspecific variation within each species - see post 168. As such, your test does not assist in determining the veracity of the method. Furthermore, as I've stated multiple times, we already know that the fundamental assumption of the method - species can be accurately delimited by color alone - is flawed. This means that the method, for this particular application, isn't going to function well.

Edited July 9, 2015 by Arete

[StringJunky]

 

.... we already know that the fundamental assumption of the method - species can be accurately delimited by color alone - is flawed. This means that the method, for this particular application, isn't going to function well.

...and 170-odd posts in, they blindly soldier on.

[tar]

String Junky,

 

The fact that species cannot be accurately delimited by color alone was decided upon years ago. Possibly before 24bit color.

 

I am just checking Arete's claim based on WT's theory, a smart phone app I just got the other day, and the undercolor reduction method newly described, to come up with an ACTUAL R/G number for the bands on bees that Arete brought to our attention as being of the same color, and as examples of why speices can not be IDed by color. If we can successfully read the ACTUAL R/G number off of bee 3 and off of bee 8 and they are different numbers, then we can take the test to the field and see if a different picture of each species, 3 and 8, yield likewise, different identifiable numbers in that band.

 

A reasonable test of Whatever Theory's claims against Arete's claims. One claim or the other will have to be adjusted, based on the outcome of the challenge.

 

I take offense at you saying I am blindly soldiering on, in the face of scientific fact already established. I am trying to falsify WT's claim.

Is Arete's claim somehow immune from falsification, just because nobody has yet falsified it?

 

Regards, TAR