From: Terry W. Colvin (fortean1@mindspring.com)
Date: Sun Jan 12 2003 - 23:28:08 MST
The first thing to say is that human vision, from cornea through optic
nerve is extraordinarily complex. Second, while I read about everything
published on vision during graduate school, over the last 3 decades my
reading in the area has been hit-and-miss.
Young (1801) and Helmholtz (1852) postulated a tri-chromatic theory of
color vision, that humans have 3 types of rods. Helmholtz drew theoretical
curves of sensitivity to the various wavelengths for each . . . the middle
one a "bell curve" maximally sensitive to green and the two others rather
skewed toward the central wavelengths and maximally sensitive to red/orange
and to blue/violet. The theory was strongly supported at the retinal level
in 1964 when Marks, Dobelle & MacNichol reported in _Science_ that the
absorption curves of single monkey and human cones fell into 3 groups.
[They extracted single cones and passed various wavelengths of light
through it "the long way" . . . important both because the photochemical
molecules are aligned for maximum reaction to light from that direction and
because of the limited sensitivity of the instrument that measured how much
came out the other end, i.e., was not absorbed. To appreciate the
difficulty of dealing with a single cone, consider that each human eye
contains 6 to 7 million cones and 75 to 150 million rods.] While their
data showed that the cones maximally sensitive to blue and to green were
more similar than Helmholtz had proposed a century earlier, most accepted
the idea of 3 types of cones, differing in their relative sensitivity to
the various wavelengths.
Thaddeus Cowan <tc.pw@Relia.Net> pointed to Cornsweet's statement
that there is a fourth sensitivity spectrum in the normal human eye, that
of the rods.
Quite correct, but generally considered irrelevant to color vision.
The clearest data demonstrating that the rods play no role in the
perception of color comes from comparing what humans see when a small
colored stimulus is focused on the dead center of the eye (fovea centralis,
which contains no rods) and when the same stimulus is focused sufficiently
off center so that it hits many rods (including about 20 degrees off
center, where the greatest rod concentration is found and rods are more
numerous than cones) . . . there is no difference in color naming / color
perception by many different measures, hence no impact of the rods.
Unfortunately for direct demonstration of this point, and vision research
generally, there is no place on the retina where only rods occur.
While it is possible that rods play some role in the perception of
large colored stimuli, which have been rarely studied, most vision
researchers consider that unlikely . . . even though the neural paths
leading from rods and cones are not completely separate even within the
retina much less in the optic nerve. Incidentally, rods and cones are
distinguished either by differential histologic staining reactions or
absorption distributions . . . their name was assigned before we found that
the cones in the fovea are rod shaped.
With 3 types of cones (the only receptors generally considered to be
involved in color vision) the typical human is trichromatic . . . and also
has rods. The term tetrachromatic is reserved for those rare individuals
thought to have four types of cones (plus rods).
While it has long been recognized that rods react differently to the
various wavelengths [for example, the amount of 510 nanometer or
"blue-green" light that can be detected by the rods under the most ideal
conditions is minuscule in contrast with the amount of light of around 680
nm or "orange" required before it is seen; both appear gray]. But this
fact is primarily detected under special laboratory conditions whereas
every day we name colors . . . hence we focus on the color-naming measure.
One set of figures at hand shows 6% of males and 0.4% of females as
Deuteranopes (i.e., weak or nonfunctioning "green" cones). Since we can't
yet directly measure the retina without destroying vision, the most precise
measures of color weakness come from asking people to match colors.
Deuteranopes with milder problems simply require more green than typical to
match yellow with a mixture of red and green. In its stronger forms what
most of us see as bluish-red, green, and gray are said to be the same color
(i.e., perfect matches).
Protanopes (2% of males and 0.04% of females, with "red" cone
problems) range from those requiring more red than typical to match yellow
with a combination of red and green to those who report red, bluish-green,
and gray to be the same color.
The extremely rare Tritanope (0.0001% of males, too few females to
calculate) range from requiring more blue than typical to match blue-green
with a mixture of blue and green to a handful who report purplish blue,
greenish yellow, and gray as identical.
Historically the names trichromat, dichromat, and monochromat come
from a slightly different color matching task. Consider the "color
circle," formed by twisting the linear spectrum of wavelengths into a
circle so that hues directly opposite each other, when mixed in about equal
proportions, are seen as gray [i.e., green and red yield gray; blue and
yellow yield gray . . . reference here to the "additive mixing" of shining
two separate lights on the same surface, not to the "subtractive mixing"
produced by two color filters in front of a single bulb or when mixing
paint].
How many different wavelengths, spaced reasonably apart on the circle,
are required for a human observer to match all possible single wavelengths
of light?
** an individual who has no cones will only require one . . . the
entire world appears gray, the only variation in vision is brightness;
hence the monochromat. One can argue whether "1" color is seen or "0", but
historically "mono" has been used.
** a few individuals are satisfied to match all possible hues when
only 2 wavelengths fairly far apart on the color circle are mixed.
** most humans require only three lights to make a good match for all
possible wavelengths.
** I have been told (but never seen in a refereed journal or standard
text) of a few individuals, all women, who are not satisfied with the
matches unless four colors from the circle are used; presumably these are
the tetrachromats referred to by Emlyn [sorry, I've lost the fuller
attribution]. While color weakness affects roughly 16 times more men than
women, some of the most interesting variations are found in females.
Another one, also with only a handful of cases, consist of humans with
normal color vision in one eye and marked color weakness in the other eye.
One set of figures shows 92.0% of males and 99.6% of females to be
trichromats. Anomalous trichromats, those presumed to have "weakness" in
one of their cone types, constitute 5.9% of males and 0.4% of females.
Dichromats, those with only 2 functioning cone types, make up 2.1% of males
and only 0.03% of females. Rod monochromats, those with cones either
absent or not functioning . . . and the only ones mentioned here that are
truly "color blind", are 0.003% of males and 0.002% of females.
If everything seems clear so far, watch out. The major rival to the
Young-Helmholtz tri-chromatic theory of color vision was Hering's 1874
opponent-process theory. Following in the tradition of astute observations
by Leonardo da Vinci and Goethe, he postulated four basic colors (in
mutually antagonistic pairs) and three receptors:
* a red--green receptor, red from breakdown or dissimilation and green
from assimilation
* a yellow--blue receptor, the yellow being dissimilation
* a white--black receptor, the white being dissimilation
Until the single cone sensitivity data of Marks et al in the mid-1960s
there was no clear winner between the two theories, as some color vision
effects supported one theory, other effects its rival. For a short period
after Marks et al it appeared that Hering was wrong. But then the vision
theorists started asking why several phenomena so strongly supported the
Hering notion. For example, the negative after-image:- stare at a vivid
red patch for a minute then immediately look at a white or light gray . . .
the gray appears greenish (and vice versa); clearly an opponent-process
sort of event.
After some additional work the currently accepted understanding
arose:- the Young-Helmholtz tri-chromatic theory describes how the retina
works (i.e., there really are 3 types of cones in the typical human eye)
but the neural pulses are converted into something very much like Hering's
opponent-process for the passage through the optic nerve to the brain.
Hence color vision is a combination of a tri-chromatic physiology (retinal)
and a two-part opponent-process physiology (neural). Depending on the
task, one of these may play the major role and the other be almost
undetectable . . . hence the age-old inability to decide between the two
theories.
And then we have, at about the same time as Marks et al, the Nobel
prize-winning work of Hubel and Wiesel, who showed that some visual
receptors more readily detect vertical bars of light whereas others are
more sensitive to horizontal or 90-degree angle bars. How does this tie in
to color vision and the Young-Helmholtz / Hering notions? And we haven't
even mentioned moving stimuli . . . .
Charles
---------------------
[...]
> With 3 types of cones (the only receptors generally considered to be
> involved in color vision) the typical human is trichromatic . . . and also
> has rods. The term tetrachromatic is reserved for those rare individuals
> thought to have four types of cones (plus rods).
The only reference I could find for "tetrachromatic" was in Cornsweet's
book. Even though I've taught this stuff for 30+ years, it would not
surprise me to find a few lacunae in my knowledge. Here is your
opportunity to teach an old dog a new trick. I would like to know the
following:
1) What retinal area is sensitive to the "second" interpretation of
green? The retinal areas sensitive to red and green are small and
overlap (pretty much). The areas sensitive to blue and yellow are
larger an overlap (pretty much). Does the retinal area for the new
color enjoyed by the tetrachromat overlap the red-green area or fall
somewhere between the blue-yellow area?
2) What does the spectral absorption curve for the fourth cone look
like?
3) What is the new color's complement?
These are very basic questions. Surely these data are available
somewhere.
Finally, how does the existence of a fourth cone change our
understanding of the opposing process mechanisms in the LGN? Normally,
for the trichromat there are three centers--a blue-yellow, red-green,
and black-white (light present-light absent). Is there a fourth center
for "new green" and its complement?
Thad Cowan
--
Terry W. Colvin, Sierra Vista, Arizona (USA) < fortean1@mindspring.com >
Alternate: < fortean1@msn.com >
Home Page: < http://www.geocities.com/Area51/Stargate/8958/index.html >
Sites: * Fortean Times * Mystic's Haven * TLCB *
U.S. Message Text Formatting (USMTF) Program
------------
Member: Thailand-Laos-Cambodia Brotherhood (TLCB) Mailing List
TLCB Web Site: < http://www.tlc-brotherhood.org >[Vietnam veterans,
Allies, CIA/NSA, and "steenkeen" contractors are welcome.]
This archive was generated by hypermail 2.1.5 : Wed Jan 15 2003 - 17:35:51 MST