Equine vision compares horses' sight to that of a human with red-green colour blindness. It also says that a horse has sensors in the blue wavelength region and in the green wavelength region.

Does this mean that horses do not sense red wavelength light or that they sense it as if it were (dull?) green light?

Is human red-green blindness one of the red or green sensors not working or is it that both sensors feed the same signal analysis paths (not sure what the right terminology is here).

? -- SGBailey (talk) 13:34, 5 October 2016 (UTC)[reply]

For humans, see Red–green color blindness. In humans it can be caused by either lack of the red cones, or lack of the green cones, or deficiency of one of these. The horse article says that horses have green and blue cones (ie, lacking red) so that would most closely correspond to Protanopia in humans. SpinningSpark 13:57, 5 October 2016 (UTC)[reply]

• I assume you have some familiarity with the relation between wavelength and color (or are willing to read the somewhat technical link). If you immediately understand the figure to the right, don't bother with the following wall of text.

Human color vision is based on three types of rod cells cone cells, which are light receptors. Each of those converts the light spectrum coming from a direction into a single measure (number), so it sees in "black-and-white". However, the details of how the light intensity is converted into an electric signal along the neuron vary for each (honestly I have no idea how/why, but it probably is similar to the way a filter works); as a consequence, the sensitivity varies with the wavelength (i.e., some incident monochromatic light will trigger a higher or lower neuron signal depending of the wavelength, for the same energy).

Imagine now that you look at an object that is "blue", i.e. your eye is struck by a flux of photons of wavelength 420nm (give or take). All the receptors will get their share of photons, but only the S-cones (the "blue" ones, S standing for short wavelength) will yield a high signal, since the others have a low sensitivity at those wavelengths. Your brain, upon receiving low signal from the "red" and "green" wires, and a high signal from the "blue" wire, will perceive a blue image.

Notice also that the figure I included has a fourth type of receptor, "R" for rod cells. I do not know why we do not see in four colors thanks to it, but it is active at much lower light intensities, so it could be "saturated" (damn, how is there nothing I want on the DAB page for saturation?) during daylight and/or the brain wires it on the same as "blue". It is responsible for the Purkinje effect.

So, horses see two colors instead of the human's three, but it is misleading to say those are "blue and green" as (in all likelihood) the horse's S and M cones have a different absorption spectrum as human's. Objects that the horse will perceive as most luminous will be seen as blue or green by humans, but "blue" and "green" are anthropocentric concepts: in Reality (tm), there is an infinity of wavelengths and they are all equal.

If you ever wondered why the visible spectrum has a very short area of yellow (compared to other colors), the reason is that the M and L cones' perception peaks are closer together than the S and M cones'. The yellow part is where there is approximatively as much of "green" as there is of "red" and that happens on a short range of wavelengths. Tigraan^{Click here to contact me} 16:14, 5 October 2016 (UTC)[reply]

@Tigraan: You stated above "Human color vision is based on three types of rod cells, which are light receptors". Colour vision is based on cone cells, not rod cells. I suggest you edit this. DrChrissy ^(talk) 18:06, 5 October 2016 (UTC)[reply]

Obviously we don't know what color qualia animals have, but there are rare humans with normal color vision in one eye and no functioning L/M cones in the other eye (unilateral dichromacy), and those people report that they see blue and yellow in the dichromatic eye, so it seems plausible that dichromatic mammals also see those two hues.

Note that dichromats see only two hues (probably blue and yellow), but trichromats (human ones at least) see a continuous circle of hues, not just three. -- BenRG (talk) 22:25, 5 October 2016 (UTC)[reply]

That "dichromats see only two hues" is a little misleading way of stating this. They see a range of hues which can be represented by a linear function of different proportions of response from the two cone types. That is, a one dimensional colour space. A linear function won't do for trichromats, a two-dimensional colour space is needed to fully represent the colours seen by trichromats such as humans. The colour circle spoken about by BenRG is merely the outer boundary of this colour space. Species with even greater numbers of cone types (that's right, humans do not have the best colour vision, in fact it is rather poor) will require proportionately higher dimensioned colour spaces to represent them. SpinningSpark 22:53, 5 October 2016 (UTC)[reply]

So what creatures have 4 or more sets of colour sensors? -- SGBailey (talk) 09:56, 6 October 2016 (UTC)[reply]

The amazing mantis shrimp! (beat our own resident mantis to it :-)---Sluzzelin talk 10:49, 6 October 2016 (UTC)[reply]

Yes, the mantis shrimp has 16 cone types, the highest known number of any species. Four of these are used to detect differing polarizations of light (something humans cannot detect at all) rather than colours, but against that, the mantis shrimp has a system of filters in front of the cones which only allow through a narrow band of wavelengths. This effectively greatly multiplies the number of primary colours the mantis shrimp can distinguish. More generally, many other crustaceans have good colour vision and many birds and insects have four cone types. SpinningSpark 15:20, 6 October 2016 (UTC)[reply]

Most mammals aside from primates have two cone types. The L and M cone pigment genes in primates are next to each other on the X chromosome and are thought to be duplicates of the single L/M pigment of other mammals which later drifted away from each other. There is a lot of variation in the L/M cone peak sensitivity wavelength across different mammal species; I don't know where you'd find information about horses specifically. The L/M cones, regardless of peak wavelength, are sensitive to some degree across the whole visual spectrum, so variation in the peak wavelength doesn't have a large effect on what wavelengths you can see. But humans with no L cones do have a harder time seeing red than humans with no M cones (it simply looks darker). This article has more information about color vision in mammals. -- BenRG (talk) 22:25, 5 October 2016 (UTC)[reply]

The OP asks: “Does this mean that horses do not sense red wavelength light or that they sense it as if it were (dull?) green light?” When hunting deer, it is wise (and legally required in some places) to wear safety orange camouflage fatigues so that other hunters can see you clearly and don't pump a 56 round into you. So deer probably see no difference between orange and the green of vegetation. It also seem pretty well accepted now, that the dichromatic vision of reindeer (and other critters) can also see into the UV. Power Lines Look Like Terrifying Bursts of Light to Animals. Don't know about horses but I would imagine, that because their night vision is so excellent, that their rods are sensitive to UV also, because star-light (like sunlight) is rich in it, but but cones are not sensitive to this end of the spectrum, so the rods excel compared with our own at night. Also, think that is misleading anthropomorphism to compare horses vision with human colour blindness. From the horse's point of view, she may consider us humans as having defective vision. --Aspro (talk) 17:50, 6 October 2016 (UTC)[reply]

Cones are sensitive to UV. See for example[1][2] DrChrissy ^(talk) 18:32, 6 October 2016 (UTC)[reply]

Aspro is correct in warning us about being anthropomorphic and anthropocentric. For example, the graph at the top of this thread shows the frequency responses of the 3 human cone types. Look at the response curve for the blue receptor - it extends beyond 400nm which means we have the photoreceptors to actually be sensitive to UV. However, our lens does not transmit UV so it means we don't perceive it. A good reference is here[3]. My understanding is that the addition or loss of a cone type has a additive effect, so gaining or losing a cone skews all colours, it does not simply add or lose the range of colours the cone is sensitive to. So, asking if a horse (dichromat) sees green as perceived by humans (trichromats) is an anthropocentric question. DrChrissy ^(talk) 19:10, 6 October 2016 (UTC)[reply]

I found an interesting article from 1996 here which explains:

"In general, nonprimate mammals are dichromats. They have S (blue)-cones and one type of M/L-cone, whose exact spectral sensitivity in the red/green range varies according to the species (Jacobs, 1993). Our immunocytochemical data show that the horse has at least two cone types (Fig. 5), and a behavioral study suggests that it is a dichromat (Pick et al., 3994)."

This points me to an article [4] that says "The subject could reliably discriminate blue (462 nm) vs. gray, and red (700 nm) vs. gray without regard to reflectance (P<0.001), but could not discriminate green (496 nm) vs. gray." A one-subject study is less than impressive as behavioral studies go, but searching for references to 1996 article gets me to [5], which cites another oddly underpowered study here claiming that horses could discriminate red and blue vs. gray reliably, but some could and some couldn't learn to tell yellow and green from gray. There's something a little mysterious going on in the middle, as you'd expect from a dichromat. The two color peaks are surprisingly close together (to me, but apparently other ungulates are similar): "Values for the spectral peaks of these photopigments have been estimated at 545 nanometres (nm) [59, 72, 75]) and 429 nm [72]." So it is possible for them to tell color in one direction or the other by which photopigment is closer, and I imagine the precise shapes of those curves are really important for deciding hue. The situation might be really different from the curve above - you see humans with red-green color blindness are getting basically no help at all from the blue cones. The horses are more like humans telling a whole bunch of hues from red to green and beyond by comparing the exact readouts from the red and green receptors (except of course they're using S and M/L that way, not M and L). Wnt (talk) 13:12, 7 October 2016 (UTC)[reply]

Just to be pedantic, I would like to point out that not all human females are thought (by some researchers) to be simply trichromatic.The women with super human vision Some are thought to exhibit tetrachromacy, (which is mayby why – to give them the benefit of the doubt -I may have found some of them them such a pain to go shopping with for cloths etc...). On reflection, I agree with DrChrissy that it is probable the blue cones that enable people, that have undergone cataract removal to perceive such wonderful blues. I wonder if their spectacles are made from polycarbonate, as normal glass blocks UV, and the thick lenses need after this procedure would be heavy if made from glass rather than plastic. Richard Feynman is claimed to be the only person to see the Trinity atomic test explosion without the goggles provided, relying on a truck windshield to screen out harmful ultraviolet wavelengths. [[6]]. Any ophthalmologists out there?--Aspro (talk) 13:52, 7 October 2016 (UTC)[reply]

Erm, actually I just reread the quote I used above and looked at the figure and realize it's not as different from the human as I thought - I must have misread a number the first time around. The M/L receptor is a bit more green than red, and the blue receptor is just a tiny bit closer. Wnt (talk) 19:04, 8 October 2016 (UTC)[reply]

I connected a 220uF capacitor, a 8.1k resistor and a diode in series to the mains. The voltage was 224V. I measured 255 VDC across the capacitor. It wouldn't charge to any higher voltage. I don't really understand the result. Where does the number 255 V come from (why not, say, 224V*sqrt(2)/2=158 V?) Asmrulz (talk) 18:14, 5 October 2016 (UTC)[reply]

In case it is not clear: do not do this.

Electrocution is easy, painful, and nearly instantaneous.

When you are electrocuted, it is actually advisable for your rescuers not to touch you, and to leave you in writhing, agonizing pain, until you are dead, because if they touch you to rescue you, they will also die.

Stop playing with mains electricity.

Nimur (talk) 18:42, 5 October 2016 (UTC)[reply]

You appear to be saying that the mains voltage is 224 VAC, in which case this half-wave rectification circuit should produce a voltage across the capacitor of very close to 1.4 x 224 = 313.6 VDC. The probable reason why it doesn't reach that figure is that the capacitor has leakage resistance that is significant in proportion to the 8.1K resistor. The voltage across the 8.1k is 313.6 - 255 = 58.6 VDC. According to Ohms Law, the voltage divides inversely across the two resistors. Therefore, the leakage resistance R_L of the capacitor is 8100 x 255/58.6 = 35247.4 ohms or about 35.247k. It's pretty leaky. This amount of leakage indicates it would conduct about 8 mA at the full 313.6 VDC. Akld guy (talk) 18:55, 5 October 2016 (UTC) Edited to correct error of scale. Akld guy (talk) 19:01, 5 October 2016 (UTC)[reply]

Or, your devices are behaving non-ideally, because you're passing mains current through them, and the capacitor is shorting out, and the resistor is shorting out, and the high voltage is leaping from conductor to conductor, energizing the insides- and outsides- of your wires, because even at the low voltage of a hundred volts, passive components don't behave like they do in textbooks. Stop playing with dangerous voltages. You can be killed by a voltage at fractions of the levels you're measuring. Nimur (talk) 19:21, 5 October 2016 (UTC)[reply]

What on earth are you talking about? Do you know anything about electronics? So long as the capacitor is voltage rated to 300 VDC or so, no damage can occur at the stated voltage levels unless the capacitor is faulty. In fact, the circuit described is an appropriate way to test the performance of the capacitor, since the 8.1k resistor limits any destructive effect if the capacitor happens to be short circuited. The OP has asked a reasonable technical question about something he does not understand, and the way he has posed the question implies that he knows quite a bit about the components and how they should behave. I have been involved with the design and testing of power supplies up to 1,000 volts DC for about 50 years and am currently building and testing several power supplies of up to 600 VDC for use with my WW2 military surplus radio transmitters AN/ARC-5. Akld guy (talk) 20:25, 5 October 2016 (UTC)[reply]

Here is a reprint of an article hosted at the Ohio State University: Strange as it may seem, most fatal electric shocks happen to people who should know better. Do not post advice to Wikipedia's reference desk recommending anybody plays with high voltage power supplies, or insinuating that you are magically impervious to high voltage "because you're a smart person." Your years of experience and all your expertise do not change the electrical conductivity of your innards. Nimur (talk) 20:44, 5 October 2016 (UTC)[reply]

cool, thanks. why should it be 313.6 V, though? Isn't the voltage on the live wire "centered" around 0? This would mean the instantaneous voltage at any time is ±Vpp/2, the diode is forward-biased for one half-cycle and reverse-biased for another, the capacitor never sees the full amplitude Asmrulz (talk) 20:20, 5 October 2016 (UTC)[reply]

I assumed European mains voltage since you stated 224 V RMS. Is that not correct? At 224 V RMS, the capacitor should charge to the peak voltage, close to 1.4 X 224 = 313.6 VDC. It's not centered around 0. Akld guy (talk) 20:36, 5 October 2016 (UTC)[reply]

that's correct. Nominally, it's 230. I think it drops in the evening. I still don't get the half-rectifier thing. Would the capacitor charge to Vpp/2 (~160V) if it was centered? Does it mean there's a DC offset? Asmrulz (talk) 20:45, 5 October 2016 (UTC)[reply]

Nimur: absolutely, thank you. Don't try this at home, kids etc, etc (or at least use an isolating transformer, hehe) The OP is qualified because he apprenticed in an area roughly to do with electricity Asmrulz (talk) 20:28, 5 October 2016 (UTC)[reply]

No, even if you "know what you are doing," don't go hooking up components to mains electricity. That is very stupid, and if you learned anything during your "apprenticeship," you would have learned not to do that. Nimur (talk) 20:45, 5 October 2016 (UTC)[reply]

They didn't say anything about hooking up random stuff to mains. Just about the VDE safety rules, types of cabling wrt ampacity, fuses, RCDs, earthing systems and that's only the electrical stuff. The scare quotes are just nasty. I don't understand what the problem is. Would you react differently if this was AllAboutCircuits or another such site? Asmrulz (talk) 21:25, 5 October 2016 (UTC)[reply]

Your mains voltage was 224 v RMS. At that value, its peak-to-peak voltage is 2 x 1.414 x 224 = 633.47 V. The peak voltage is one-half of that, or 316.7 V. My 313.6 was a close approximation due to rounding down to 1.4. Effectively your single diode is passing half the waveform to the capacitor, which should charge to almost the full peak voltage 316.7 (about 0.6V is lost across the diode), but in this case the capacitor seems to be significantly leaky w.r.t the 8.1k series resistor, so the capacitor is not charging to that full peak voltage. Akld guy (talk) 21:27, 5 October 2016 (UTC)[reply]

oh... you're right. if rms=230V=A/1.4, then A=230*1.4=325V, then Vpp=A*sin(wt)=-325..+325=650Vpp. I wasn't aware of that. So it is centered? Asmrulz (talk) 21:53, 5 October 2016 (UTC)[reply]

No, it's not centered, at least as I understand you to mean centered. It's based on the zero crossing point of the waveform, since the diode passes only one half of the waveform (that's why it's called a half-wave rectifier circuit). The diode passes only the positive half or negative half (above or below the 0 line) depending which way it's connected. So the waveform passed to the capacitor is not a waveform based around the 0 line. It's based on it as the reference level. Akld guy (talk) 22:02, 5 October 2016 (UTC) Edited slightly for clarity. Akld guy (talk) 22:07, 5 October 2016 (UTC)[reply]

I just mean whatever is on the live wire, is centered wrt neutral. So the current does actually reverse direction, otherwise there'd be 650-0.6V across the cap Asmrulz (talk) 22:10, 5 October 2016 (UTC)[reply]

@Asmrulz I made this checklist that may clarify your setup and apologise if some points seem banal.

1. . In your circuit that connects to the mains voltage, all metal parts are securely supported, and insulated or very carefully kept out of your touch. YES/NO?
2. . You are aware that a capacitor charged to hundreds of volts is a shock hazard even after the charging source is removed. YES/NO?
3. . Your mains supply is a regular installed household 50 or 60 Hz AC grid that is sinusoidal. YES/NO? (Some heavy equipment can distort the supply waveform, some battery inverters give only a square wave or approximate sinewave output. This might explain the anomaly. An oscilloscope would be needed to give a definite answer.)
4. . You measure the AC input voltage on an RMS reading AC voltmeter. YES/NO? (This is the usual calibration of a Multimeter which actually responds to the average voltage magnitude. True RMS AC meters are more expensive and will give the same readings if the waveform is sinusoidal.)
5. . You measure the voltage across the capacitor using a voltmeter on a DC (not AC) range. The voltmeter has high enough input resistance to have negligible loading effect on your circuit. YES/NO? (A suitable voltmeter could have a 100 µA movement, a quoted sensitivity of 10,000 ohms/volt and therefore 5 megohm input resistance on the 500V range.)
6. . The diode is a silicon type with a reverse voltage rating of at least 350V 650V YES/NO? Common suitable diodes are 1N4004, 1N4005, 1N4006 or 1N4007. Lower numbers in this series are unsuitable.
7. . Are you sure the resistor is 8.1 kilohm? YES/NO? I ask because it is not a common preferred value while the value 8.2 kilohm marked grey-red-red-silver is extremely common.
8. . In your series circuit, is the 220µF capacitor en electrolytic type with its + terminal connected to the diode cathode (marked with a band) YES/NO? (The resistor is in series with both and may be between them. The point is that you must not apply a reverse voltage to the electrolytic.)
9. . The 220µF electrolytic is clearly marked with a maximum voltage rating of at least 350V YES/NO?
10. . You are aware that the combination of 8.1 kilohm and 220µF creates a slow Time constant equal to (8.1E3 x 220E-6 = 1.78 seconds) which means that the capacitor takes this much time to charge to 63% of its final voltage. YES/NO?

TWO SIMPLE TESTS

1. . Never forgetting warnings 1. and 2. above, there should be no detectable heating of any of the 3 components when the circuit is left energized for several minutes. YES/NO?
2. . Reversing the mains input does not change the measurement. YES/NO? Any difference means the AC supply is unbalanced with a distorted waveform or a residual DC voltage. AllBestFaith (talk) 00:53, 6 October 2016 (UTC)[reply]

I agree with every point on the checklist made by AllBestFaith, with the exception of point 6. The diode's reverse voltage rating MUST be at least 650 V PIV, because during the non-conducting part of the AC cycle, the diode is subjected to the voltage in the capacitor (which should be 325 VDC at full mains voltage of 230V RMS) plus the peak value of the opposite polarity voltage during that non-conducting portion (325 V peak). A PIV rating of at least 1,000 V is suitable. The better the reliability as the rating goes higher. For me, the OP's question was simply about the phenomenon he observed, and I assumed he knew the function and ratings of the components and what he was doing, otherwise he wouldn't have asked the question. Akld guy (talk) 02:38, 6 October 2016 (UTC)[reply]

Thank you that's a good point. An inadequate PIV rated diode could even explain the anomalous measurement. I have edited point 6 accordingly. AllBestFaith (talk) 11:25, 6 October 2016 (UTC)[reply]

Thank you for changing the figure. An inadequately rated diode could not cause the observed phenomenon, unless it had failed due to that inadequate rating. Failure in a diode with inadequate PIV rating typically results in the diode becoming short circuited. The capacitor is presumably an electrolytic type and this type is likely to explode if wrong polarity voltage is applied to it, especially if the voltage source is a low impedance supply such as the mains supply described here. The OP hasn't reported any such disastrous behaviour, so we can probably assume the diode is OK and the capacitor is simply leaky. I once could have lost an eye when I reverse-connected an amplifier I had built to my car battery. An electrolytic capacitor exploded its contents out one end like a bullet and hit the garage door with considerable force. Be careful. Akld guy (talk) 20:40, 6 October 2016 (UTC)[reply]

AllBestFaith, yes to all, more or less (no exposed parts, correct DMM ranges, no heat (I didn't run it for minutes, though, but certainly longer than several RC), sufficient cap voltage rating, correct polarity of the cap wrt diode.) The resistor is indeed coded as 8k2 but it measures 8k1. There was a small difference depending on the order of the parts (255 V with the cap in the middle, 250 V with the diode in the middle.) I don't know if it means anything. At what end the live wire was, made no difference with either setup Asmrulz (talk) 21:38, 6 October 2016 (UTC)[reply]

I don't know what the diode is. I tried it variously with a big black diode with thick leads from a monitor and a generic black diode from a CFL Asmrulz (talk) 21:52, 6 October 2016 (UTC)[reply]

Since another diode is available a test will answer the reverse breakdown question.

3. .Add a second diode in series with the series circuit. Increased DC voltage across the capacitor indicates that reverse breakdown current was flowing in the single diode.

The RC product in point 10 would be the exponential charging time if the mains supplied DC. However on the AC mains the capacitor charges only in half-frequency pulses that become progressively shorter. Wait a minute or more for the capacitor to charge fully before measuring its DC voltage.

The small difference when you change the order of parts might be your DMM having difficulty measuring DC consistently with both terminals floating at AC potential. Are you using a mains isolation transformer? An isolation transformer will let you connect any chosen point of the circuit, such as the cap negative, to ground.

The transformer should have its own mains fuse, 1 Amp is adequate. AllBestFaith (talk) 22:51, 6 October 2016 (UTC)[reply]

Thanks, AllBestFaith. Are you saying reverse breakdown voltages add with every additional diode? I'm not using an isolation transformer. The DMM runs on a battery (A23), is it still susceptible to the floating terminals thing (i.e. this isn't about ground loops)? Asmrulz (talk) 11:36, 7 October 2016 (UTC)[reply]

Reverse breakdown voltage for 2 diodes in series is more than either diode alone (which is enough for test no. 3) but do not rely on the PIV ratings exactly adding. The datasheet for 1N4001..1N4007 diodes has a Fig. 4 Typical Reverse Characteristics that shows the onset of reverse current is abrupt and very temperature dependant. Reverse current/voltage behaviour is properly characterized for Zener diodes which we are not using while manufacturers of power rectifier diodes do only a pass/fail test of PIV. So they can legitimately set up a production line for 1N4007 diodes and mark them all as 1N4001, etc.

Your battery powered DMM is probably not screened against external electric field and you might find that the reading changes slightly when you change its position or orientation. The only remaining test I shall suggest is:

4. .Time the decay in DC voltage across the capacitor after mains is disconnected. The seconds to decay to 37% of the start voltage divided by 220E-6 gives the effective leakage resistance of the capacitor and DMM combined in parallel. This calculation can't be very accurate (because electrolytic capacitors have inexact capacitances and no reason to obey Ohm's law, and the DMM has updating delays) but it may explain your missing volts. AllBestFaith (talk) 20:04, 7 October 2016 (UTC)[reply]

Any basin and time period. Sagittarian Milky Way (talk) 21:17, 5 October 2016 (UTC)[reply]

I started to wonder why or how painted bat got such a coloration, given that they rely on echolocation rather than vision. This book, for example, says that "most bats have somber coloration" and "even the Asian painted bat... the most colorful species... is difficult to see in the tropical forest foliage of Southeast Asia", so presumably it's not aposematism. A random mutation that got stuck? Brandmeister^talk 21:50, 5 October 2016 (UTC)[reply]

According to this it may be from their diet of plantain. Our article suggests that they are insect eaters, but that is not referenced at all. The book I cited is old - 1894, but nevertheless appears a reliable source. Matt Deres (talk) 01:12, 6 October 2016 (UTC)[reply]

To clarify, that book is claiming that the coloration served the purpose of camouflage (i.e. crypsis in the animal world), so as to blend in to the fruits while eating. It appears that older work made an erroneous assumption that they at fruit, because they were often found on fruit. This [7] 2105 study says they like to roost on banana leaves during the day, but that they mostly feed on spiders. This [8] also supports that the family is insectivorous.

The camouflage story indeed makes sense, but it would be rather difficult to rigorously defend such an assertion by ruling out random fluctuation and luck. In principle, they could dye a bunch of bats black and see if they suffer a higher predation rate. I've seen some studies like that, but (unsurprisingly) cannot find one on this bat. SemanticMantis (talk) 14:33, 6 October 2016 (UTC)[reply]

My apologies; I thought that link would update as I read forward. If you follow the link and flip over the next page (p287) it says "...Indian painted bat, which feeds on plantains, which, when ripe, are of a bright yellow or orange colour, speckled with black, thus almost exactly similar to the bat." However, I am not terribly surprised they are insectivorous - I just couldn't find anything quickly to confirm that before retiring for the night. To be honest, when I first saw a picture of the bat, my first thought was "I wonder what bug they eat that's giving them that complexion?" and was a little puzzled by the assertion that they ate fruit. But, when in doubt, go with the reference... Matt Deres (talk) 15:54, 6 October 2016 (UTC)[reply]

I did read that, but I think the writer is saying that it blends in with the fruit, not that the fruit gives it its color. I can see now how other readings are defensible, but across the animal world, coloring to blend in with surroundings is much more common than color drawn from diet. SemanticMantis (talk) 18:05, 6 October 2016 (UTC)[reply]

Our article also suggests it's camoflauge although it's unsourced. Nil Einne (talk) 04:42, 7 October 2016 (UTC)[reply]

No question, but there are notable instances where the food provides the colour. Matt Deres (talk) 14:08, 7 October 2016 (UTC)[reply]