Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
~2~989/9 1$3.00+ 0.00 YiftbnRes.Vol. 31,No. 3, pp. 567-576,1991 Printedin Great Britain.All rightsn~erved Copyright Q 1991Rrgamon Press plc ULTRAVIOLET PHOTORECEPTION IN CARP: MICROSPECTROPHOTOMETRY AND BEHAVIORALLY DETERMINED ACTION SPECTRA CRAIG W. HAWRYSHYN’* and FERENC I. HAROSI~*~ ‘Division of Biological Sciences, Section of Ecology and Systematics, Corson Hall, Cornell University, Ithaca, NY 14853, *L&oratory of Sensory Physiology, Marine Biological Laboratory, Woods Hole, MA 02543 and 3~~~ent of Physiology, Boston University School of Medicine, Boston, MA 02118, U.S.A. (Receiued 3 March 1990; in revised form 10 July 1990) Abstract-This study demonstrates correlations between U.V. sensitivity and microspectrophotometric absorption spectra determined sequentially for the same group of individuals. We used the heart-rate ~n~tioning technique to measure spectral sensitivity of carp, a species known to have u.v.-sensitive photor~pto~, Mean spectral sensitivity (n = 3) determined with a spectrally-broad background (450 nm long pass filter) revealed a small but consistent U.V. peak (.I,, of 380 nm) in addition to the other long wavelength peaks. An intense blue-green background (490 nm) produced a more prominent U.V.peak (,I,, of 4OOnm) when a 450 nm longpass filter was added to the background. Microspcctrophotometric measurements of u.v.-sensitive photoreceptors from one individual, which belonged to the group used in the spectral sensitivity experiments, revealed an average ,I,, of 377.5 nm (SD +4.5 nm, n = 5 ceils). Bfeaching and dichroic measurements of these receptors ensured that we were examining typical vertebrate visual pigments and not stable photoproducts. The mean spectral sensitivity points were compared with the U.V.and blue-sensitive visual pigment absorption spectra. A linear subtractive model and ocular media absorption were used in this comparison for the various photic conditions used in the heart-rate conditioning experiments. The model successfully described the sensitivity of the test fish in two cases but in a third case there was some discrepancy. The model generated curve was broader than the spectral sensitivity of the u.v.-sensitive cone mechanism on the shortwave side even though the ocular media corrections had been accounted for. Ultraviolet photoreception Ocular media transmission Microspectrometry Cone photoreceptors Heart-rate conditioning Spectral sensitivity INTRODUC’MON Most vertebrate visual pigments can absorb ultraviolet (u.v.) radiation through /I-band absorption. It is now evident, however, that some vertebrates also possess u.v.-sensitive visual pigments with a-band absorption in the U.V. spectrum (Harosi & Hashimoto, 1983; Avery, Bowmaker, Djagmoz & Downing, 1983; Harosi, 1985; Bowmaker & Kunz, 1987). The spectra1 nature of quanta1 absorption by the u.v.-sensitive visual pigments is in~uen~d by the presence of ocular media which may or may not transmit U.V. radiation (Hawryshyn, Chou & Beauchamp, 1985). Species occupying photic environments hosting high ambient intensities of U.V. radiation, usually possess U.V. ocular *Present address, to which reprint requests should be addressed: Department of Biology, University of Victoria, P.O. Box 1700, Victoria, B.C., Canada V8W 2Y2. Carp (Cyprinus carpio) filters. Species with u.v.-filtering media may incur several selective advantages including protection from photic damage to photoreceptors, and the reduction of excessive scattering and chromatic aberration that could degrade image contrast (Chou & Hawryshyn, 1987). Conversely, fish with more u.v.-transparent ocular media (e.g. goldfish and rainbow trout) may receive adequate levels of U.V. radiation for photoreception. However, individual variability in ocular media transmission can impose a variable degree of narrowing on the U.V. peak (Hawryshyn et al., 1985). Three different comparative surveys have shown that U.V. photosensitivity is especially prevalent in birds (Chen & Goldsmith, 19861, and in fish (Fukurotani & Hashimoto, 1984; Harosi (8 Fukurotani, 1986). Several technical approaches have been used to explore U.V. photosensitivity in vertebrates: (1) MSP (microspectrometry) of single cones has identified 567 568 CRAIG W. HAWRYSHYN and FERENC I. HAROSI u.v.-sensitive cone receptors in cyprinids (Avery 1983; et al., 1983; Harosi & Hashimoto, Harosi, 1985; Bowmaker & Kunz, 1987) and cyprinidonts (Harosi & Fukurotani, 1986); (2) ERG (electrophysiology) was used to measure U.V. sensitivity in birds (Chen, Collins & Goldsmith, 1984; Chen & Goldsmith, 1986); (3) classical conditioning of behavior (Kreithen & Eisner, 1978; Hawryshyn & Beauchamp, 1982, 1985; Hawryshyn, Arnold, Chaisson & Martin, 1989); (4) operant conditioning of behavior was used to measure U.V. sensitivity and wavelength discrimination in the near u.v.-spectrum in fish and birds (Parrish, Ptacek & Kevin, 1984; Neumeyer, 1985, 1986; Goldsmith, 1980; Douglas, 1986). For many vertebrate species, the identity of photoreceptors mediating U.V. photosensitivity is unclear. For example, goldfish exhibit a well defined U.V. peak in its spectral sensitivity (Hawryshyn & Beauchamp, 1985) and perform well on near U.V. wavelength discrimination tasks (Neumeyer, 1986), yet a u.v.-sensitive cone photoreceptor remains unidentified. This may be due to miniature long single cones and miniature short single cones eluding the recording beam. These cones remain the favored candidates for the u.v.-sensitive receptors in goldfish. MSP (Avery et al., 1983) and the use of operant conditioning to examine spectral sensitivity (Douglas, 1986) revealed U.V. sensitivity and u.v.-sensitive cone pigments in another cyprinid species, the roach. However, the full absorbance spectrum and dichroic properties are unknown parameters of the u.v.-sensitive cones. In a recent report by Harosi and !Fukurotani (1986), a good correlation was demonstrated between cone absorbances and the response of horizontal cells (u.v.-sensitive triphasic or tetraphasic) in several species of fish. In the present study, we measured the spectral sensitivity and the absorbance spectrum of u.v.sensitive cones of carp from the same group. Furthermore, measurements of the ocular media transmission permitted a direct comparison of the u.v.-sensitive cone absorbance spectrum with the spectral sensitivity. MATERIALSAND METHODS Heart -rate conditioning experiments Animals. Seven carp (Cyprinus carpio) obtained from SP Engineering Technology (Salem, Mass.) were used. Initially, three additional carp were trained but these individuals lost their conditioned responses at the beginning or during the spectral sensitivity experiments. The three that lost their conditioned responses were discarded from the test group. They ranged in size from 13.0 to 17.2 cm and 48.5 to 114.0 g. All fish were kept in the holding facility at 20°C on a 12HL/12HD cycle for at least a month prior to conditioning. Fish were fed on alternate days a mixed diet of live meal worms, frozen brine shrimp and pellets. Immobilization procedure and set -up of jish. Fish were immobilized during each conditioning and experimental session to prevent changes in the position of the retinal image of the presented stimuli, Each fish was first anesthetized with sulfonate (MS-222), at tricaine methane 250 mg/l (immersion 2-5 min) and then immobilized with d-tubocurarine chloride (0.45 ,ug/g body weight) injected into the dorsal musculature at several sites. The animal was held in a Plexiglass restrainer, placed in a test tank (80 l), and artificially irrigated with a 3 ml/set flow of aerated water (all procedures and care of fish were in compliance with the Cornell University Council on Animal Care). The position of the restrainer was adjusted so that the right eye of the fish was 5 cm from the center of a circular Albanene (Kuffler and Esser Co.) back projection screen (9 cm dia.) and the pupillary plane of the eye was parallel with respect to the back projection screen (see Fig. 1). Stimuli. A three-channel optical system was set up so that the light could be superimposed from two background channels and a stimulus channel on the back projection screen. The wavelength of the stimuli was controlled by 22 narrow-band (10 nm at half-max bandwidth) interference filters (Pornfret) spanning a 320-740 nm range. The transmission propertiesof the filters were measured using a Perkin-Elmer spectrophotometer. Short- or long-pass cutoff filters (Corion) were used in combination with the interference filters to correct for transmission leaks outside the primary band-pass if necessary. Irradiance measurements were made for each stimulus filter combination with an International Light Radiometer (Detector SEE 400 1360 calibrated by International Light Inc.) with its probe positioned inside the test tank at the center of the back projection screen. The stimulus intensity was controlled by a u.v.-grade Inconel-coated (Kodak) neutral density wedge which was calibrated in 0.2 log unit steps for each wavelength setting used during these U.V. sensitivity in carp 569 Fig. 1. Top view of the experimental apparatus. AR = artificial respirator; BCH 1 = background channel 1; BCH2 = background channel 2; c = condensing lens; D = diaphragm; EKG = electrocardiogram electrode and leads; ES =electronic shutter; FL = field lens; FR = fish restrainer; FT = filter tray; H = light housing; IFW = interference filter wheel; NDW = neutral density wedge; PL = projection lens system; PW = Pyrex optical window; S = light source, 250 W tungsten lamp; SC = stimulus channel; SE = shock electrodes and leads; SM = photodiode stimulus monitor system; TF = test fish carp (Cvprinus c&o); TT = test tank (flat black). experiments. Additional u.v.-grade Inconelcoated (Corion) neutral density filters were used in the stimulus channel as needed. Stimulus duration was controlled by an electronic shutter and was kept constant at 4sec during these experiments. A 250 W tungsten lamp (EHJ Spectra Lamps) powered by a Hewlett-Packard 24 V d.c., 10 A regulated power supply, provided the light for the stimulus channel. The background channels also contained 250 W tungsten lamps and were powered by 24 V d.c., 5 A regulated power supplies. The intensities of the background channels were controlled by the same type of neutral density filters as the stimulus channel and their spectral composition was controlled by the same type of short- and long-pass cutoff filters that were used to correct for transmission leaks in the stimulus channel. The absorption of the water between the back projection screen and the fish’s eye was determined by measuring the absorption of 1 cm of the water using a Perkin-Elmer spectrophotometer. Conditioning protocol. After being immobilized and positioned in the test tank, the fish were trained using the heart-rate conditioning technique (Beauchamp & Rowe, 1977; Hawryshyn & Beauchamp, 1985). The purpose of training was to elicit a conditioned decrease of heart rate in response to detection of a light stimulus. This was accomplished by presenting a light stimulus to the fish, followed immediately by a low level (2-3 mA, 3-4 V r.m.s.) electric shock delivered to the caudal peduncle. The fish’s heart rate was monitored on a Grass Polygraph using two-pin electrodes inserted just distal to the pectoral fins. The response criterion was chosen to be a stimulus period heart beat interval that was 1.5 times the length of the average pre-stimulus period (10 set) heartbeat interval (see Fig. 2). During the conditioning sessions, a moderate intensity white background (for spectral characteristics of the backgrounds see Fig. 5, right panel) was used, the stimulus size was kept constant at 2.5 cm and the distance between the fish’s eye and the back projection screen was a constant 1Ocm. The conditioning protocol consisted of repeatedly presenting a bright 600 nm stimulus until the fish gave five consecutive criterion responses and then presenting a series of stimuli of different wavelengths and intensities to achieve appropriate levels of generalization. Each wavelength and intensity combination was repeatedly presented until two consecutive criterion responses were given. At the end of each conditioning and experimental session three blanks were presented to assure that responses were related to CRAIGW. HAWRYSHYN and FERENCI. HAR~X.I 570 Fig. 2. Conditioned EKG responses (arrows) of one fish to 340 nm stimuli. Threshold was determined using a single ascending series of stimulus radiances. Each successive stimulus was 0.2 log units more intense than the last. Three consecutive trials in a series are illustrated: top trace no. 1, subthreshold stimulus; middle trace no. 2, threshold stimulus; bottom trace no. 3, suprathreshold stimulus. Threshold was always preceded by at least three consecutive stimulus radiances of the series not giving a response, and followed by at least one giving a response. A conditioned response was defined as an interbeat interval during the stimulus which was at least 1.5 times the average interbeat interval in the 10 set preceding the stimulus. Note: after training, shock followed only those trials on which criterion responses were measured (after Hawryshyn & Reauchamp, 1985). the light stimulus, and not to extraneous factors such as sound or vibration. Threshold determination. Threshold at a particular wavelength-background condition was determined by presenting a subthreshold stimulus intensity followed by increasing intensity steps (0.2 log unit) until the test fish gave two consecutive positive responses. Threshold intensity was the first intensity level to which the fish responded. To minimize the incidence of false positives, the threshold intensity, the first positive response (i.e. criterion bradycardia), had to be preceeded by at least three negative responses (three intensity increments). Spectral sensitivity experiments determined the threshold of detection of a stimulus at wavelengths varying from 320 to 740 nm for each test fish. Test wavelength thresholds were determined in random order to reduce any effects of time of stimulus presentation on the resultant spectral sensitivity curve. Microspectrophotometric Spectrophotometer. trophotometer measurements The dichroic microspec(DMSP) described elsewhere (Harosi & MacNichol, 1974; Harosi, 1982a,b) was used in these experiments. It is a singlebeam photometer that simultaneously records average and polarized transmitted fluxes as a function of wavelength. Spectral recordings. The cross-section of the measuring light was a rectangle of ca 1 x 3 pm in the specimen plane. The preparation, consisting of a quartz sandwich, was placed on the sliding-gliding stage of the DMSP, and in it transversely oriented photoreceptors were located under dim red background illumination. The sample transmittance was usually recorded in eight scans, whereas reference transmittance was recorded in 8 or 16 scans through an adjacent cell-free area (for technical details, see Harosi, 1987). RESULTS MSP absorbance spectra Absorbance spectra were measured from cones located in the central regions of the carp retina. Measurements shown in Fig. 3 were from one individual (fish C-10 used in the spectral sensitivity experiments). Except for the absence U.V. sensitivity in carp (A) UV-sensitive 1.1 cones - 0.6 k I; 300 + + + +> +*+~&~++++++~+3+~+ I I , 350 400 450 I I I I I I 1 I I I I 550 600 650 300 350 400 450 500 550 600 650 600 650 ++, 500 0 (B) Blue-sensitive cones 1.1 1.0 0.9 0.6 0.7 0.6 0.5 0.4 _.. 300 350 400 450 500 550 400 Wavelength 450 500 5% 600 650 (nm) Fig. 3. Absorbance spectra of cone photoreceptors in test carp C-10. (A) Ultraviolet-sensitive cones: average of five single-cell prebleach absorption spectra. The solid line (in curves A-D) represents the Fourier-smoothed spectrum (for technique, see Harosi, 1987). (B) Blue-sensitive cones: average of five single-cell prebleach absorbance spectra. (C) Green-sensitive cones: prebleach absorbance spectrum of one cell. (D) Red-sensitive cones: average of three single-cell prebleach absorbance spectra. of miniature cones, photoreceptors in the carp preparations were similar to those found in the goldfish @tell dz Harosi, 1976). The u.v.-sensitive and blue-sensitive visual pigments were located in short single cones. These cones were indistinguishable from one another as viewed in the light microscope. The numerical distribution of the cone types examined is shown in Table 1. Because the microspectrophotometer was adjusted to give optimal results at short wavelengths, the u.v.- and blue-sensitive cones Table 1. Visual pigments found in the cones of test carp C-IO” Cone type u.v.-sensitive Blue-sensitive Green-sensitive Red-sensitive Number of cells found Number of cells averaged Peak of Fouriersmoothed curves A,, (nm) 10 22 2 5 5 5 377.5 458.0 532.0 600.0 I 2 ‘Note our search for photoreceptors was very much biased toward short single cones and thus the frequency of occurrence of the pigment type should not be construed as being representative of the natural distribution. yielded the clearest spectra. The green-sensitive absorbance spectrum was somewhat noisier, but we feel it to be acceptable since its I,,,,, compared favorably to previous estimates (J,,, 535 nm, Harosi, 1985). However, recordings made on the red-sensitive cones showed deviations from the expected 620 nm I,,, . We have provided this absorbance spectrum to demonstrate the presence of red-sensitive cones in this individual; in agreement with what is expected for carp. An average absorbance spectrum from five u.v.-sensitive cones had a high amplitude a-band peaking at 377.5 nm (Fig. 3). Dichroism in u.v.-receptors We performed transverse linear dichroism measurements on u.v.-sensitive cones in the bleached and unbleached states in order to test the hypothesis that the u.v.-sensitive cones contain typical vertebrate visual pigments rather than stable photoproducts (Nolte & Brown, 1972) or sensitizing pigments (Kirschfeld, CRAIGW. HAWRYSHYN and FERENCI. 512 Franceschini & Minke, 1977) as seen in insects. A high amplitude anisotropic response in the cl-band of the unbleached u.v.sensitive cones with a mean dichroic ratio of 2.8 (n = 3 cells) was observed (Fig. 4A). When the u.v.-sensitive cones were bleached by exposure to mon~hromatic u.v., the dichroic ratio was reduced close to unity (mean of 1.2, n = 3 cells, Fig. 4B). Clearly, these data for transverse measurements indicate that u.v.-sensitive cones share the dichroic and bleaching properties of vertebrate photoreceptors. Note, however, that these transverse measurements do not test for the presence of axial dichroism in cones nor do they give any clues about the biophysical arrangements involved in the orthogonal polarization sensitivity that was recendy documented in teleost cone mechanisms (Hawryshyn & McFarland, 1987). Spectral sensitivity The spectral sensitivity of carp was measured under three different background field conditions (curves in Fig. 5 arbitrarily displaced). The different background adapting fields used in these experiments were expected to produce a 6 g 3 004 0.03 lB’ Bleached I -0,021 1 ’ ’ ’ ’ ’ ’ ’ 300’ 350 400 450 500 550 600 650 A Wavelength (nm 1 Fig. 4. Linear dichroism in ultraviolet-sensitive cones in test carp C-10. (A) Mean log,, linear dichroism of three unbleached u.v.-sensitive cones with a A,,,,, of 376.5 nm and a dichroic ratio of 2.8. (B) Mean log,, linear dichroism of three u.v.-sensitive cones after exposure to bleaching illuminations ,I,,,_ 400.0 nm and dichtoic ratio 1.2. Note that the consequence of bleaching is the loss of transverse Iinear dichroism. HAROSI I I(A) 6-“cb;”, ,- ,a 400 500 600 700 Wovelength 400 500 600 700 (nm) Fig. 5. Spectral sensitivtty of carp measured with different chromatic backgrounds. Mean spectral sensitivity of several indi~duals is shown on the left panel and relative power spectra of the chromatic background is shown on the right panel. Left panel-curve A: mean (1 SEM) spectral sensitivity (n =4 carp C-IO, C-II, C-14, C-IS) using a 4SOnm longpass plus 490 nm (lo nm half-max bandwidth) backgrounds (see right panel curve A). Curve B: mean spectral sensitivity (n = 3 carp C-l, C-4, C-S) using a 450 nm long pass background (see right panel curve B). Lines in curve B have the same representation as curve A. Curve C: mean spectral sensitivity (a = 2 carp C-14, C-15) using a white background (see right panel curve C). Righrpanel+curve A. background 1, a 250 W tungsten source operating at S A with a 450 nm longpass filter and a 2.0 neutral density filter; background 2, a 250 W tungsten sour= operating at 5 A with a 490 nm (IO nm half-max bandwidth) and a 1.0 neutral density filter. Curve B: background I, a 250 W tungsten source operating a S A with a 450 nm longpass filter and a 2.0 neutral density filter. Curve C: background 1, a 250 W tungsten source operating at S A with a 2.2 neutral density filter. variable degree of light adaptation in the u.v.sensitive cone mechanism. The mean spectral sensitivity curves (of several individuals) are shown on the left panel of Fig. 5 and the chromatic background on the right panel. Shortwave sensitivity varied with changes in the intensity of U.V. radiation in the background. When U.V. radiation was eliminated with the 450 nm longpass filter, shortwave sensitivity increased. In addition, the narrow-band bluegreen background (490 nm, IO nm half-max bandwidth) used in curve A appeared to produce the best defined U.V.peak. The addition of the blue background reduces the sensitivity of the blue-sensitive mechanism thus revealing the U.V. peak. However, the U.V. peak was also somewhat higher than the other peaks in curve A compared to curve B. 573 U.V.sensitivity in carp Ocular media transmission Transmission of light through the ocular media of the carp eye was measured with an integrating sphere spectrophotometric technique (Hawryshyn et al., 1985; Chou & Hawryshyn, 1987). This technique measured the transmission of the anterior part of the eye, including the cornea, lens and some vitreous humor. Figure 6 shows the average absorbance of the five carp eyes (derived from different animals used in the spectral sensitivity determinations). Corrections for ocular media transmission were made to absorption spectra of cone pigments and illustrated in Fig. 7 along with the spectral sensitivity points taken from Fig. 5. Correlation of spectral sensitivity pigment absorption spectra with visual 1.4 r z 0.6- 0 z z 06- u 04 - 350 400 450 Wavelength 500 (nm 550 . -* Figure 7 illustrates the three corrected spectral sensitivity curves from Fig. 5. The output of cone photoreceptors depends on a number of principal parameters such as: absorption spectrum of the cone mechanism, spectral transmission of the ocular media, and neural interactions, especially between spectrally opponent cone mechanisms. Many examples of this type of interaction can be seen in electrophysiological and behavioral studies examining red-green opponency in fish. Some recent evidence indicates that there may be an interaction between the u.v.- and blue-sensitive cone types in several cyprinid species. Wavelength discrimination experiments in goldfish (Neumeyer, 1985, 1986) have shown that there is good discrimination in the shortwave part of the 6 I t 600 650 1 Fig. 6. Spectral absorbance of carp ocular media. The solid line represents the average ocular media absorbance of five carp eyes (carp C- 11, C- 14, C- 15)measured with an integrating sphere technique (refer to Hawryshyn et al., 1985). .’ 1 LOQ Unit IA . 400 I I , 500 600 700 Wavelength . (nm) Fig. 7. Correlation of ocular media corrected spectral sensitivity and visual pigment absorption spectra. The data for curves A-C were taken from Fig. 5. Curve A: the triangles represent the mean sensitivity points (this is also the case for curves B and C) while the dashed line represents the u.v.- and blue-sensitive cone absorption spectrum corrected for ocular media absorption generated by the linear subtractive model (see Results section for the k coefficients used in the linear subtractive model). Curve B: the dashed lines for the u.v.- and blue-sensitive cone mechanisms were generated using the linear subtractive model (see Results section for k coefficients). Curve C: the dashed line for the blue-sensitive cone mechanism was generated using the linear subtractive model (see the Results section for the k coefficients). spectrum indicating that the u.v.-sensitive cone mechanism is an active participant in wavelength discrimination. Such performance depends on the interaction of at least two cone types. Furthermore, several studies have demonstrated the presence of u.v.-sensitive tetraphasic horizontal cells in a variety of cyprinid species (Hashimoto, Harosi, Ueki & Fukurotani, 1988) although this has yet to be shown for carp. This may be an important consideration for fitting absorption spectra to the u.v.- and blue-sensitive components of the spectral sensitivity curve. It is conceivable that opponent interactions may exist between the u.v.- and blue-sensitive cone mechanisms and that they may play a role in shaping the spectral sensitivity of fish in the shortwave part of the spectrum. A linear subtractive model was used to produce absorption spectra (or cone output spectra) to fit narrowed peaks of the u.v.- and 514 CRAIGW. HAWRYSHYN and FERENC I. HAROSI blue-sensitive cone mechanisms (Sperling & Harwerth, 1971; Douglas, 1986; Neumeyer, 1984). We suggest the hypothesis that inhibitory interactions occur between the u.v.- and blue-sensitive cone mechanism that produce narrowed spectral sensitivity peaks. A linear subtractive model for receptor opponency, comparable to that used by Sperling and Harwerth (1971), was employed to more fully understand the deviation of the action spectra from the u.v.-sensitive visual pigment absorption spectrum. The inhibitory effect of one cone mechanism on another (e.g. a u.v./b opponency) can be described in the following expressions: S,, = t (k3 VPb - k4 VP, c ); where S,,. and S, are the sensitivities of the u.v.and blue-sensitive cone mechanisms, VP,, and VPb represent the absorption of the u.v.- and blue-sensitive cones respectively, taken from Fig. 3; t is the ocular media transmittance; and k,, k,, k3 and k4 are the strength constants. In curve A, where shortwave narrowing of the u.v.sensitive cone mechanism is evident, the linear subtractive model was successful in generating a U.V. peak that was narrowed on the shortwave limb. As the coefficient k, was increased relative to k, the inhibitory effect was greatest on the longwave limb. This is due to an asymmetry in overlap of the blue-sensitive pigment spectrum on the cc-band of the u.v.-sensitive pigment (85% at 410 nm vs 78% at 340 nm). The dashed line in Fig. 7A illustrates the best fit of modelled absorption spectra (k, = 1.0, k, = 0.3, k3 = 0.25, k, = 0.25) to the spectral sensitivity points. Although the model fits well on the longwave side of the u.v.-sensitive cone mechanism there was a consistent discrepancy between ocular media corrected absorption spectra generated by the linear subtractive model and the spectral sensitivity. This observation did not appear to be related to u.v./b opponency or ocular media absorption. Shortwave narrowing has been observed in several species and laboratories and it appears to be an important problem which requires further study. The model was successful in generating curves that approximated the spectral sensitivity in curve B and curve C. A fit was observed for curve B when the coefficients were set to k, = 0.9, k2 = 1.0, k, = 0.27 and k, = 0.27. Similarly, a fit was also observed for curve C when the coefficients were set to k, = 0.9, k, = 1.3, k, = 1.0 and k, = 0.63. While both u.v.- and blue-sensitive mechanisms were evident in curve B, the u.v.-sensitive mechanism appeared to be absent in curve C, this curve being dominated by the blue-sensitive cone absorption points. This undoubtedly results from a more intense U.V. content in the background used for determining the spectral sensitivity illustrated in curve C. DISCUSSION This study relates the u.v.-sensitive cone absorption spectrum to the U.V.spectral sensitivity of carp (Fig. 7). We initially compared the ocular media-corrected u.v.-sensitive cone absorption spectrum with spectral sensitivity (a test of cone mechanism independence); however, this revealed narrowing between the mechanisms that reflected shortwave opponent interactions. This coupled with previous work suggests that the linear subtractive model could provide a more realistic correlate of sensitivity. The linear subtractive model generated curve correlates well with the u.v.- and blue-sensitive mechanism spectral sensitivity data (Fig. 7 curve B). In Fig. 7C, only the blue-sensitive mechanism is apparent, although the model does require a degree of inhibitory input from the u.v.-sensitive mechanism to produce a reasonable correlation to the data. Collectively, these data suggest the presence of a possible u.v./b interaction, as evidenced by the narrowing between cone mechanisms. Recent research has revealed the presence of u.v.-sensitive triphasic and tetraphasic horizontal cells. These cells have been reported in some cyprinid species (Harosi & Fukurotani, 1986), although not yet in carp. A curious lack of correlation was observed between the u.v.-sensitive absorption spectrum (ocular media corrected) and the U.V. peak action spectrum on the shortwave side of the curve (Fig. 7A). Following correction of the U.V. absorption spectrum for ocular media transmission shortwave narrowing of the U.V. sensitivity was evident. This was especially the case when the blue-sensitive mechanism was light adapted (Fig. 7A). This suggests that besides light adaptation of the blue-sensitive mechanism, other processes may be occurring that augment the sensitivity in the U.V. This is not a new finding since data from Chen and Goldsmith (1986) on the house sparrow show shortwave narrowing of the U.V. mechanism in U.V. sensitivity in carp comparison with a I,,, of a hypothetical 370 nm visual pigment absorption curve. A similar observation has also been made in Tiger Salamander rods (Cornwall, MacNichol & Fein, 1984). As indicated previously, our initial thoughts were that the shortwave narrowing of the u.v.-sensitive mechanism could conceivably be due to inhibitory interaction between the u.v.- and blue-sensitive mechanism in the shortwave spectrum. However, the linear subtractive model, over a wide range of k constants, could not successfully demonstrate a correlation on the shortwave side. Paradoxically, the shortwave narrowing occurs only when the monochromatic blue field (490 nm, 10 nm half-max bandwidth) is added to the 450 nm longpass field. The blue-sensitive mechanism decreases in sensitivity while the u.v.-sensitive mechanism increases, suggesting possible nonlinearity in the opponent interactions. It is not clear what causes this shortwave narrowing when the blue-mechanism is light adapted. Most of the research on U.V. photoreception in vertebrates has concentrated on passerine birds and cyprinid fishes. It is not known whether amphibians, reptiles or mammalian species also possess U.V.photoreception. Recent evidence by Hawryshyn et al. (1989) and Bowmaker and Kunz (1987) has shown the possible existence of an ontogenetic loss of U.V. photosensitivity or cones in salmonids. This observation of developmental losses of U.V. sensitivity will further complicate our attempts to survey the incidence of U.V.photoreception in vertebrates. Because of this, our search for U.V. photoreception should focus primarily on the early life history of fish. The results of this paper along with those on the horizontal cell recordings (Fukuortani & Hashimoto, 1984; Harosi & Fukurotani, 1986; Hashimoto et al., 1988) collectively suggest a possible u.v./b opponency in cyprinids and probably other vertebrate species. The precise spectral nature of the opponency is likely to be variable from one species to the next since there is variance in the A,,,,, of the U.V. mechanism. Species such as the Japanese date have two shortwave cones with a A,,,,, of 350-370 nm and 405-415 nm, clearly differing from carp which possesses a A,,,,,of 380 nm and of 455 nm. Horizontal cells that are u.v.-sensitive are either tetraphasic showing differential response to U.V. and blue stimuli or triphasic that do not show differential responses to U.V. and blue stimuli. Presumably, the possession of tetraphasic cells (not yet seen in carp) would 515 augment the potential for wavelength discrimination in the U.V. spectrum. Data from Neumeyer (1985, 1986) shows that cyprinids have good discrimination in the shortwave spectrum. Shortwave discrimination may permit intraspecific communication (Harosi, 1985, photographic plates). It may also aid in the detection of food items in the environment such as zooplankton for fish or flowers and berries for birds (see Burkhardt, 1982 for discussion). Acknowledgements-Thanks to Margaret Arnold for technical assistance in this project. Thanks to Daryl Parkyn and Luc Beaudet for reading the manuscript. This research was supported by USPHS grant EY04876 to F.I.H., NATO Postdoctoral Fellowship from the Natural Sciences and Engineering Research Council of Canada to C.W.H. and by a Cornell University Hatch Grant 403 to Dr W. N. McFarland for equipment support. REFERENCES Avery, J. A., Bowmaker, J. K., Djamgoz, M. B. A. & Downing, J. E. G. (1983). Ultraviolet sensitive receptors in freshwater fish. Journal of Physiology, London, 334, 23-24P. Beauchamp, R. D. & Rowe, J. S. (1977). Goldfish spectral sensitivity: A conditioned heart-rate measure in restrained or curarized fish. Vision Research, 17, 617-624. Bowmaker, J. K. KcKunz, Y. (1987). Ultraviolet receptors, tetrachromatic colour vision and retinal mosaics in the brown trout (Sahno trutfa): Age-dependent changes. Vision Research, 27, 2101-2108. Burkhardt, D. (1982). Birds, berries and u.v.: A note on some consequences of U.V. vision in birds. Naturwissenschaften, 69, 153-157. Chen, D. M. & Goldsmith, T. H. (1986). Four spectral classes of cone in the retina of birds. Journal of Cornpararioe Physiology A, 159, 473-479. Chen, D. M., Collins, J. S. & Goldsmith, T. H. (1984). The ultraviolet receptor of bird retinas. Science, 285, 337-340. Chou, R. B. & Hawryshyn, C. W. (1987). Spectral transmittance of the ocular media of the bluegill (Lepomis macrochirus). 1214-1217. Canadian Journal of Zoology, 65, Cornwall, M. C., MacNichol, E. F. Jr & Fein, A. (1984). Absorbance and spectral sensitivity measurements of rod photoreceptors of the tiger salamander, Ambystoma figrinum. Vision Research, 24, 1651-1659. Douglas, R. H. (1986). Photopic spectral sensitivity of a teleost fish, the roach (Rutilus rutilus). with special reference to its ultraviolet sensitivity. Journal of Comparatiue Physiology A, 159, 4155421. Fukurotani, K. & Hashimoto, Y. (1984). A new type of S-potential in the retina of cyprinid fish: The tetraphasic spectral response. Investigative Ophthalmology and Visual Science (Suppl.), 24, 118. Goldsmith, T. H. (1980). Hummingbirds see near ultraviolet light. Science, 207, 786788. Harosi, F. 1. (1982a). Recent results from single-cell microspectrophotometry: Cone pigments of frog, fish and monkey. Color Research and Applications, 7, 135-141. 576 CRAIG W. HAWRYSHVNand FERENCI. HAROSI Harosi, F. I. (1982b). Phase modulated microspectrophotometry of photoreceptors. Investigative Ophthalmology and Visual Science (Suppl.), 22, 189. Harosi, F. I. (1985). Ultraviolet- and violet-absorbing vertebrate visual pigments: Dichrotic and bleaching properties. In Fein, A. & Levine, J. S. (Eds.), The visual system. New York: Liss. Harosi, F. I. (1987). Cynomolgus and rhesus monkey visual pigments. Application of fourier transform smoothing and statistical techniques to the determination of spectral parameters. Journal ofGeneralPhysiology, 89, 717-743. Harosi, F. I. & Fukurotani, K. (1986). Correlation between cone absorbance and horizontal cell response from 300 to 7OOnm in fish. Investigative Ophthalmology and Visual Science (Suppl.), 27, 192. Harosi, F. I. & Hashimoto, Y. (1983). Ultraviolet visual pigment in a vertebrate: A tetrachromatic cone system in the date. Science, 2.22, 1021-1023. Harosi, F. I. & MacNichol, E. F. Jr (1974). Dichroic microspectrophotometer: A computer assisted, rapid, wavelength-scanning photometer for measuring linear dichroism in single cells. Journal of the Optical Society of America, 64, 903-9 18. Hashimoto, Y., Harosi, F. I., Ueki, K. & Fukurotani, K.-K. (1988). Ultra-violet-sensitive cones in the color-coding systems of cyprinid retinas. Neuroscience Research (Suppl.), 8, S81-S95. Hawryshyn, C. W. & Beauchamp, R. D. (1982). Aberrant high blue sensitivity in goldfish. Investigative Ophthalmology and Visual Science (Suppl.), 22, 282. Hawryshyn, C. W. & Beauchamp, R. D. (1985). Ultraviolet photosensitivity in goldfish: An independent U.V. retinal mechanism. Vision Research, 25, 1I-20. Hawryshyn, C. W. & McFarland, W. N. (1987). Cone photoreceptor mechanisms and the detection of polarized light in fish. Journal of Comparative Physiology A, 160, 459-465. Hawryshyn, C. W., Chou, B. R. & Beauchamp, R. D. (1985). Ultraviolet transmission by the ocular media of goldfish: Implications for ultraviolet photosensitivity in fishes. Canadian Journal of Zoology, 63, 1244-l 25 I. Hawryshyn, C. W., Arnold, M. G., Chiasson, D. J. & Martin, P. C. (1989). The ontogeny of ultraviolet photosensitivity in rainbow trout (Salmo gairdneri). Visual Neuroscience, 2, 247-254. Kirschfeld, K., Franceschini, N. & Minke, B. (1977). Evtdence for a sensitising pigment in fly photoreceptors. Nature, London, 269, 386390. Kreithen, M. L. & Eisner, T. (1978). Ultraviolet light detection by the homing pigeon. Nature, London, 272, 347-348. Neumeyer, C. (1984). On the spectral sensitivity of goldfish: Evidence for neural interactions between different *‘cone mechanisms”. Vision Research, 24, 1123-l I3 1. Neumeyer, C. (1985). An ultraviolet receptor as a fourth receptor type in goldfish color vision. Naturwissenschaften, 72, 162. Neumeyer, C. (1986). Wavelength discrimination in the goldfish. Journal of Comparative Physiology A, 158, 203-213. Nolte, J. & Brown, J. E. (1972). Ultraviolet-induced sensitivity to visible light in ultraviolet receptors of Limulus. Journal of General Physiology, 59, 186200. Parrish, J. W., Ptacek, J. A. & Kevin, L. W. (1984). The detection of near-ultraviolet light by nonmigratory and migratory birds. The Auk, 101, 53-58. Sperling, H. G. & Harwerth, R. S. (1971). Red-green cone interactions in the increment-threshold spectral sensitivity of primates. Science, 172, 18tK184. Stell, W. K. & Harosi, F. I. (1976). Cone structure and visual pigment content in the retina of the goldfish. Vision Research, 16, 647-657.