F. W. MUNZ
26
between lights of different wavelengths. How many different visual pigments are involved, one for each primary color? Are they similar to other
visual pigments? Does each pigment occur within a different class of receptor cells, or may the visual pigments be mixed together? Direct evidence has recently been obtained by microspectrophotometry of individual retinal cones of fishes; this technically difficult method has also
been applied to visual pigments of frogs and primates.
Microspectrophotometry was first applied to the cones of carp by
Hanaoka and Fujimoto (1957), who reduced a beam of monochromatic
light to a diameter of 3 p, small enough to pass through the outer segment. Exposure to bright light changed the absorbance, but the difference
spectra obtained in this way were only roughly similar to those of known
visual pigments. More sensitive photomultipliers have recently reduced
the amount of bleaching caused by absorption of the measuring light
(Liebman and Entine, 1964), or the absorbance data have been subjected
to computer analysis to compensate for this bleaching (Marks, 1965).
These authors have studied the goldfish; they agree that there are three
different classes of cones, each possessing a single visual pigment. The
data of Liebman and Entine seem more amenable to direct interpretation,
but they gave no estimates of A, , , , . Marks found that the cone pigments
appeared to be generally similar to other known visual pigments, but
no good evidence with respect to the product of bleaching has yet been
obtained. The approximate A,,,,, values of these pigments are 625 nm
(red), 530 nm (green), and 455 nm (blue). Each of the pigments
occurred in single cones and each in twin cones. In more than 50 single
cones, the ratio was approximately 2 red:4 green:l blue. The two
members of a twin pair never had the same pigment. Of 30 pairs of
twins, 29 were red-green pairs and one was blue-green. No red-blue
pairs were found.
It is fair to ask, how good is the evidence that goldfish have color
vision? Electrophysiological activity in the retinae of goldfish has been
found to be consistent with this idea ( MacNichol et al., 1961; Tamura
and Niwa, 1967). Potentials have been measured from the inner segments
of single cones in the closely related carp by Tomita et al. (1967), who
compared their results with those of Marks. Behavioral tests show that
goldfish can discriminate between different colors when brightness cues
are eliminated (e.g., McCleary and Bernstein, 1959; Muntz and CronlyDillon, 1966 ) although this capacity has not been investigated systematically (see Yager, 1967). The biochemical and physiological evidence that
has been gathered at several different levels should be relevant therefore
to the problems of color vision.
26
between lights of different wavelengths. How many different visual pigments are involved, one for each primary color? Are they similar to other
visual pigments? Does each pigment occur within a different class of receptor cells, or may the visual pigments be mixed together? Direct evidence has recently been obtained by microspectrophotometry of individual retinal cones of fishes; this technically difficult method has also
been applied to visual pigments of frogs and primates.
Microspectrophotometry was first applied to the cones of carp by
Hanaoka and Fujimoto (1957), who reduced a beam of monochromatic
light to a diameter of 3 p, small enough to pass through the outer segment. Exposure to bright light changed the absorbance, but the difference
spectra obtained in this way were only roughly similar to those of known
visual pigments. More sensitive photomultipliers have recently reduced
the amount of bleaching caused by absorption of the measuring light
(Liebman and Entine, 1964), or the absorbance data have been subjected
to computer analysis to compensate for this bleaching (Marks, 1965).
These authors have studied the goldfish; they agree that there are three
different classes of cones, each possessing a single visual pigment. The
data of Liebman and Entine seem more amenable to direct interpretation,
but they gave no estimates of A, , , , . Marks found that the cone pigments
appeared to be generally similar to other known visual pigments, but
no good evidence with respect to the product of bleaching has yet been
obtained. The approximate A,,,,, values of these pigments are 625 nm
(red), 530 nm (green), and 455 nm (blue). Each of the pigments
occurred in single cones and each in twin cones. In more than 50 single
cones, the ratio was approximately 2 red:4 green:l blue. The two
members of a twin pair never had the same pigment. Of 30 pairs of
twins, 29 were red-green pairs and one was blue-green. No red-blue
pairs were found.
It is fair to ask, how good is the evidence that goldfish have color
vision? Electrophysiological activity in the retinae of goldfish has been
found to be consistent with this idea ( MacNichol et al., 1961; Tamura
and Niwa, 1967). Potentials have been measured from the inner segments
of single cones in the closely related carp by Tomita et al. (1967), who
compared their results with those of Marks. Behavioral tests show that
goldfish can discriminate between different colors when brightness cues
are eliminated (e.g., McCleary and Bernstein, 1959; Muntz and CronlyDillon, 1966 ) although this capacity has not been investigated systematically (see Yager, 1967). The biochemical and physiological evidence that
has been gathered at several different levels should be relevant therefore
to the problems of color vision.