- Email: [email protected]
Frequency response characteristics of isolated retinas from hatchling leatherback (Dermochelys coriacea L.) and loggerhead (Caretta caretta L.) sea turtles
Journal of Neuroscience Methods 178 (2009) 276–283
Contents lists available at ScienceDirect
Journal of Neuroscience Methods journal homepage: www.elsevier.com/locate/jneumeth
Frequency response characteristics of isolated retinas from hatchling leatherback (Dermochelys coriacea L.) and loggerhead (Caretta caretta L.) sea turtles Kenneth Horch a,∗ , Michael Salmon b a b
Department of Bioengineering, University of Utah, 50 S. Central Campus Dr., Salt Lake City, UT 84112, USA Department of Biological Sciences, Florida Atlantic University, Box 3091, 777 Glades Road, Boca Raton, FL 33431-0991, USA
a r t i c l e
i n f o
Article history: Received 20 August 2008 Received in revised form 12 December 2008 Accepted 12 December 2008 Keywords: Vision Electroretinogram Marine turtle Leatherback Loggerhead Frequency sensitivity
a b s t r a c t Electroretinographic recordings were made from hatchling loggerhead and leatherback sea turtle eyecup preparations during presentation of sinusoidally modulated lights of different frequencies, mean intensities and colors. Cross-correlation analysis was performed to determine the extent to which the responses followed the intensity modulated light sources. For both species mean light intensity had no significant effect on the frequency modulated responses over a 1.5 log unit range of intensities. Both species showed the best following to blue light and the poorest tracking to red light. Leatherback retinas did not follow frequencies above 10 Hz, while loggerhead responses extended out to 15 Hz. These visual low pass filter characteristics are consistent with attributes of the visual ecology of each species, as well as with the latencies and slow rise times exhibited by these retinas to brief flashes of light. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Vision plays an important role in many aspects of marine turtle behavior such as orientation (Salmon, 2003; Witherington, 1991, 1997; Witherington and Martin, 1996), habitat selection (Witherington, 1992), as well as detection of prey (Constantino and Salmon, 2003; Salmon et al., 2004) and predators (Heithaus et al., 2002). Even so, our understanding of relationships between visual physiology, ecology and behavior among marine turtles is in its early stages (Bartol and Musick, 2003). It is known that sea turtles possess multiple visual pigments (Granda, 1979) and potentially can discriminate between wavelengths that extend from the near UV to red, although this has yet to be demonstrated behaviorally (Crognale et al., 2008; Horch et al., 2008; Levenson et al., 2004; Mäthger et al., 2007). An important characteristic of visual systems is their ability to resolve rapid changes in light intensity (their visual temporal bandwidth: informally their flicker fusion frequency). Fritsches (Fritsches and Warrant, 2006; Southwood et al., 2008) has reported a surprisingly high upper frequency limit near 60 Hz for adult green turtle retinas. This is at variance with reports of others who found bandwidths at or below 20 Hz in green turtle retinas, and even lower cutoff frequencies in leatherbacks (Dermochelys coriacea L.) and loggerheads (Caretta caretta L.), during flicker photometry mea-
∗ Corresponding author. Tel.: +1 801 585 1981; fax: +1 801 585 5361. E-mail addresses: [email protected] (K. Horch), [email protected] (M. Salmon). 0165-0270/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2008.12.017
surements (Crognale et al., 2008; Eckert et al., 2006; Levenson et al., 2004). Their results are consistent with our study of retinal spectral sensitivity in hatchling leatherback and loggerhead sea turtles. We found response latencies to brief light flashes on the order of 50–100 ms, and rise times to peak of the photoreceptor (a-wave) responses on the order of 100–300 ms (Horch et al., 2008). Although the flicker photometry and spectral sensitivity work did not directly address the issue of temporal bandwidth, these results do differ from what would be expected if a 60 Hz frequency following limit were correct. We believe that a primary reason for this discrepancy is methodological and, in particular, that examining “flicker fusion frequency” using Fourier transform and related techniques is not as reliable as one would like. Our intention here is to present a more mathematically robust approach. Leatherbacks and loggerheads are of interest not only because they are endangered species for which vision plays an important role in survival, especially early during ontogeny, but because differences in their visual ecology and behavior suggest their eyes could exhibit differences in temporal resolving power. Leatherbacks as hatchlings are active primarily during the day (Wyneken and Salmon, 1992). As adults, they feed during the day and at night (Eckert and Eckert, 1989; Eckert et al., 1986; James et al., 2006) at the surface where illumination is bright (Grant and Ferrell, 1993) or by diving to depths >1000 m where down welling light is absent (Eckert and Eckert, 1989; Eckert et al., 1986). Leatherbacks during all stages of ontogeny are feeding specialists consuming slow moving gelatinous prey (jellyfishes, medusae, ctenophores, and salps) (Bjorndal, 1997).
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
Loggerheads as hatchlings are also day-active (Wyneken and Salmon, 1992). Juveniles, during their pelagic stage of development, feed primarily upon stationary prey gleaned from flotsam floating at the surface (Witherington, 2002) or found at shallow depths in the water column (Bolten, 2003). Older turtles most often return to coastal waters where they primarily consume sessile benthic prey (bivalves, soft corals, anemones), gelatinous zooplankton in the water column, and prey capable of more rapid movement (horseshoe and hermit crabs, mollusks, seahorses, portunid crabs) (Bjorndal, 1997). These ecological considerations led us to speculate that loggerheads might have somewhat “faster” eyes than leatherbacks. Additionally, most information about sea turtle vision stems from work on green turtles, with very little information available about either leatherback or loggerhead turtles at young ages (Horch et al., 2008). It is of interest to know the extent to which observations about visual physiology from one species or life stage of sea turtles can be extrapolated to other species and life stages. Both species have more than one retinal pathway (Horch et al., 2008). The existence of separate pathways might result in colordependent differences in visual temporal resolution, a concept that to the best of our knowledge has not been previously addressed in marine turtles. The claim that color effects are absent in human perceptual flicker fusion frequency is based upon the negative results obtained by Hecht and Shlaer (1936), who used themselves as subjects. The present study was designed to examine the ability of hatchling leatherback and loggerhead retinas to track changes in light intensity as a function of stimulus tempo and color, using a standard system analysis approach. This involved a mathematically well-defined stimulus (sinusoidally modulated light intensity) and a mathematically appropriate direct measure of the relationship between the stimulus and retinal response (cross-correlation analysis). Correlation analysis is a widely used and accepted method for examining information processing in the nervous system: it provides a direct measure of the relationship between two events (in this case, the stimulus and the recorded electrical response of the retina) in the form of the correlation coefficient (Levine, 1998). Additionally, it provides a direct measure of the temporal relationship between the two events (in this case, the lag or latency between the stimulus and the response). The latter is important not only for confirming that the recorded responses have latencies similar to those predicted by more direct measures of retinal response latencies (Horch et al., 2008), but also for identifying and, in appropriate circumstances, eliminating stimulus artifacts from the recordings when performing the data analysis. 2. Materials and methods Because both species are endangered worldwide, our permit required us to use no more than 6 “stragglers” of each species as subjects. Stragglers are hatchlings remaining inside the nest 72 h after the majority of the turtles have departed and are migrating offshore; turtles left behind usually do not survive, even with human intervention. All were collected during daylight hours between July and September, 2005 and 2006, from nests in Palm Beach and Broward Counties, Florida, USA. The turtles were stored in a covered Styrofoam cooler in a dark, air-conditioned laboratory. A shallow layer of moist sand or a wet paper towel was placed inside the cooler to prevent desiccation. They were sacrificed for experiments within 12 h after capture. Although the hatchlings were incapable of leaving the nest, their eyes seemed fully developed and they showed consistent responses to the visual stimuli we presented. Initial experiments were performed on two loggerheads to hone our techniques. Final experiments were done on the remaining animals permitted by our protocol (6 leatherbacks and 4 loggerheads).
277
2.1. Retinal preparations Hatchlings were anesthetized in an ice water bath for ∼20 min, or until they reached a state of torpor, decapitated at the cervical junction, and pithed according to accepted protocols (AVMA Panel on Euthanasia. American Veterinary Medical Association, 2001). The cornea, lens and vitreous humor of one eye were removed under indirect, dim illumination of a dissecting microscope (10–25×), leaving a retinal eyecup, which was filled with perfluoro FC-77 ¯ Organics) to prevent desiccation and promote oxygen diffu(Acros sion to the retina. The preparation was placed inside a Faraday cage and a Pt ground-wire was placed inside the skull. The preparation was enclosed in a clear plastic “tent”; fully humidified air mixed with 95% oxygen continuously flowed into the tent. The preparation was dark-adapted for 20 min before light stimuli were presented (Perlman et al., 1990). See Horch et al., 2008 for a more detailed description. 2.2. Recording and stimulation equipment A chloridized silver wire loop was placed in contact with the periphery of the retina. Preliminary tests confirmed that there were no responses to visual stimuli arising from the electrode itself. Signals from the retina were fed to a modified WPI DAM-50 preamplifier (gain: 1000×; bandpass: 1–1 kHz). The preamplified electroretinographic signals were monitored on a storage oscilloscope and sent to one channel of a 16-bit A/D card that allowed acquisition of the data on a laptop computer (1000 samples/s). After each stimulus sequence, the response was displayed on the computer screen. Full amplitude, sinusoidal waveforms at the desired frequency were generated by the computer and offset appropriately to provide a roughly 2.4 log unit intensity modulation. The waveform was fed to a step attenuator (attenuation range 0–2.5 log units in 0.5 log unit steps) through a digital-to-analog converter. The attenuated signal was used to control a constant current source that drove the selected LED stimulus light source. Four LEDs were used (all manufactured by Purdy Electronics Corp., Sunnyvale, CA): AND190HBA blue (430 nm peak), AND412HG blue-green (505 nm peak), AND185HOP orange (620 nm peak), and AND190HW white (significant output from 450 to 650 nm). LED spectral outputs are shown in Fig. 1. The LEDs were mounted on a carrier attached to a wafer switch so that changing from one light source to another simply required rotating the switch. This action simultaneously placed the appropriate LED at the same position relative to the eyecup and directed the driving current to it. An Ocean Optics S-2000 spectrometer was used to confirm wavelength specifications and that, due to the quantum nature of the light emission process, the color remained constant over the range of intensities used in the experiments. A UDT radiometer (model S351A, uniform response between 400 and 700 nm) was used to measure light intensity at the retina in W/cm2 . Light intensities derived from these measurements were used in all data analysis. 2.3. Stimulation and recording procedures A stimulus sequence consisted of providing a steady illumination at the mean intensity of the modulated light for 1–3 s. The duration of this steady pre-stimulus depended on the preparation and was chosen to be long enough to allow the transient response to the onset of the light to decay away. The light intensity was then varied sinusoidally (modulation depth approximately 2.4 log units) for 2 s. The signal recorded from the retina during this 2 s period was recorded on the computer, but only the last 1 s was used in the data analysis because the non-linear response properties of the
278
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
Fig. 1. Emission spectra for the LED light sources used in this study. Plots taken from manufacturer’s data sheets.
eye produced a transient response that could last up to 1 s once the modulation began. The light intensities were chosen to be bright enough to elicit photopic (i.e., cone) responses (Horch et al., 2008), but not bright enough to saturate the visual responses. Stimulus color and intensity modulation rate were presented randomly to the different preparations, but for a given color and rate, we always started with the lowest intensity and subsequently increased intensity in 0.5 log unit steps from there. Each stimulus combination was presented three times, except in cases where a visible artifact was present. In those instances additional repetitions were used to provide three “clean” records. 2.4. Data analysis Cross-correlation analysis provided a quantitative measure of the ability of the eyes to follow changes in light intensity, including accommodation of non-linearities in their stimulus–response properties, compensation for differences in recorded response amplitudes, and maintenance of temporal information about response latency. The square of the peak correlation coefficient (r2 ) is a measure of the fraction of the variability in the response that is due to variability in the stimulus (i.e., how well the system followed the stimulus). This approach is self-scaling as it measures the ability of the retina to follow intensity changes in the stimulus, largely independent of response amplitude. To find the peak correlation coefficient, the analysis was performed with different latencies (lags) between the stimulus waveform and the response waveform. From our data on response latencies to flashes of lights in these eyes (Horch et al., 2008), we expected response latencies or lags to be less than 200 ms, and so
looked at shifts between 0 and 200 ms. A very short latency would indicate a stimulus artifact, and a lag much greater than 100 ms would indicate something other than a neural response from the retina. Because the photoreceptor response to light in this preparation is negative and has a shorter time constant than the later response components (Horch et al., 2008), we expected to see a peak, negative correlation coefficient at frequencies of 10 Hz and above appearing somewhere in the vicinity of a 100 ms lag. Cross-correlation analysis was performed on the last 1 s of data using custom Matlab® scripts. The stimulus and response waveforms were first normalized to given autocorrelation values of 1. The peak negative cross-correlation coefficient was found and its square was used as the metric for how well the response followed the stimulus. Although not evident on visual observation of the responses, examination of the cross-correlation behavior revealed that there was a stimulus artifact present in the data. The properties of the artifact indicated that it was likely due to cross-talk between the stimulus waveform sent out the D/A converter and the incoming signal digitized by the A/D converter. Since the former was constant in amplitude and exhibited no lag, we were able to determine its characteristics for each recording session and generate its inverse mathematically. This inverse was added to the original recorded signal, thus canceling out the artifact, before performing the final cross-correlation analysis for the data presented here. 3. Results The top two panels in Fig. 2 show recordings and their crosscorrelograms made in response to three intensities of blue light
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
279
Fig. 2. Example of a data acquisition and analysis sequence. The top left panel shows responses from a leatherback hatchling eyecup to blue light sinusoidally modulated at 5 Hz around three different mean intensities: the lighter the trace the brighter the light. The stimulus (normalized) is presented as the trace at the bottom of the panel. The top right panel shows the cross-correlations (r values) from these recordings (calculated at different lags (latencies) between stimulus and response, uncorrected for stimulus artifact). The middle row shows the same for 7 Hz stimulation. Note that in both cases there is a decrease in response latency, indicated by a leftward shift of the peak (negative) r value, with increasing light intensity, but the peak r values are independent of response amplitude. The bottom panels are plot of the squares of the maximum cross-correlation values derived from these three light intensities over all frequencies tested with the blue LED in this animal, without (left) and with (right) artifact correction. Shading differentiates values above from those below 0.5.
modulated at 5 Hz presented to one of the leatherback eyecup preparations. The second row shows similar data for light modulated at 7 Hz. Response amplitudes decreased with increasing mean light intensity. This was probably due, either alone or in combination, to the logarithmic nature of the visual response to light intensity; to changes in light adaptation of the preparation with changes in mean level of illumination; or to increasing B-wave contribution to the response, which is opposite in sign to the photoreceptor response (Horch et al., 2008). However, the correlation values were largely the same at a given modulation rate—indicating that the eye’s ability to follow the changes in light intensity was unaffected by light intensity over this range of intensities. As seen in our earlier study, response latencies decreased with increasing stimulus intensity (Horch et al., 2008). The bottom row of the figure shows plots of the squares of the peak correlation values (without and with correction for stimulus artifact) for all frequencies tested with blue light at these three intensities in this preparation.
The left column in Fig. 3 shows three-dimensional plots of averaged response profiles for all 6 leatherbacks for each of the four colors used. On the right these data are presented in twodimensional format with standard errors. There is no consistent or statistically significant difference between the four curves for a given color. Fig. 4 shows corresponding data for four loggerhead hatchlings. As with the leatherbacks, the ability to follow light modulation did not depend on stimulus intensity. Data for each of the four stimulus intensities for each color were combined and the results are shown in Fig. 5. In both species, the frequency following curve for red light lies below the other curves (except at high rates where all the curves are approaching an asymptote and there is, essentially, no following of the stimulus for any color), an effect that is more pronounced in leatherbacks than in loggerheads. The former shows a general trend for the retinas to follow blue light better than green light better than red. However, for the loggerheads, the curves for blue and green light overlap.
280
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
Fig. 3. Average (n = 6) of squares of peak cross-correlation values as a function of light color, modulation frequency and mean intensity for retinas from leatherback hatchlings. Curves on the right, extracted from the three-dimensional plots on the left, show standard errors. The lighter the trace, the brighter the light. Data points offset for clarity.
4. Discussion Neither species appears capable of following light flicker at frequencies much above 15 Hz. Hatchling leatherback retinas respond to light modulation at 10 Hz about as well as loggerheads do at 15 Hz, supporting our suggestion that loggerhead eyes, while slow, are “faster” than leatherback eyes.
Evidence for color-dependent differences in visual temporal resolution is less obvious, but suggested by the consistently lower correlation coefficients for red light modulation than those for the other colors. These results suggest that the retinal components responding to long wavelengths in both species may be slower than those responding to shorter wavelengths. We speculate that this effect is unlikely to arise as a consequence of differences in visual
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
281
Fig. 4. Data from loggerheads (n = 4). Format as in Fig. 3.
sensitivity to red versus blue and green lights for the following reasons. First, if sensitivity was a factor, the r2 values would depend on stimulus intensity (the level of effective retinal stimulation). However, these values were independent of stimulus intensity at the intensities used here. Second, the quantal content of the bright red light was 1.2 log units higher than that of the dim blue light, a bigger difference than the difference in retinal sensitivities to the red versus the blue light (−0.9 log units for the leatherbacks, −0.2 log
units for the loggerheads) (Horch et al., 2008). Finally, the ranges of response amplitudes to red and blue lights at 10 Hz were indistinguishable from each other within a species, meaning that the measurements for different colors were effectively made at comparable response amplitudes. Such a temporal resolving power difference may occur because the longer wavelength pathways converge more to compensate for lower sensitivity to red light, at the expense of temporal coher-
282
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
of whole or intact retinal responses to changes in light intensity. In the final analysis, however, behavioral tests of the animals’ ability to perceive modulations in light intensity are needed demonstrate the extent to which animals are actually aware of and can respond to flickering lights. In conclusion, we demonstrate in this study that combining the use of sinusoidally modulated light intensity with cross-correlation analysis is a valuable method for assessing temporal resolving power in visual systems. This procedure allowed us to detect a stimulus artifact that otherwise would have provided skewed results, decreasing apparent correlation when the artifact was out of phase with the visual responses and increasing apparent correlation when the artifact was in phase with any responses or when a real visual response was absent (Fig. 2, bottom row). This method also compensates for differences in recorded signal amplitudes, provided that the responses are sufficiently large that the inherent signal averaging in the procedure can pull them out of the noise. Finally, it also provides a measure with direct physical significance: r2 is an indication of the extent to which any recorded signals are actually following the stimulus. Acknowledgments Permits to carry out this project were obtained from the U.S. Fish and Wildlife Service (TE087169-0), the Florida Fish and Wildlife Conservation Commission (TP # 173), and the Florida Atlantic University Institutional Animal Care Committee (A02-22). Support was provided by the NOAA Fisheries Sensory Biology Working Group. We thank A. McDermott for assistance with the experiments and T. Frank for comments that helped improve the manuscript. References
Fig. 5. Combined data from Figs. 3 and 4, with standard errors, for leatherback (top) and loggerhead (bottom) hatchlings.
ence in their inputs. This adaptation would be consistent with the environment at depths where these animals feed during dives, and where the visible spectrum shifts towards the blue and UV wavelengths. Note that the effects illustrated in Fig. 2 lead to two caveats. The first, and most obvious, is that the amplitude of the response, measured from whole retina preparations, is not a good indicator of frequency following by the eye. For example, the use of power spectra derived from a Fourier analysis of retinal recordings as a criterion for determining flicker fusion frequency (Fritsches et al., 2005), in addition to its theoretical and proper implementation difficulties, is basically an amplitude measure and should be employed with caution. The second, less obvious, point is that using responses to amplitude modulated light sources to estimate spectral sensitivity (Jacobs et al., 1996) may be subject to error if one does not consider that the different color pathways may show different temporal bandwidths. This is supported by the work of Crognale et al. (2008). Fig. 3 of that paper demonstrates that leatherbacks exhibit a dramatic difference in spectral sensitivity to long wavelengths depending upon the stimulation rate (8 Hz vs. 12 Hz). In addition, these two spectral sensitivities are different from one derived using an “ERG criterion” presented (Fig. 2) in the same paper. While our results differ from the significantly higher flicker fusion frequencies reported by Fritsches (Fritsches and Warrant, 2006; Southwood et al., 2008), they are consistent with related findings in green, leatherback and loggerhead turtles (Crognale et al., 2008; Eckert et al., 2006; Levenson et al., 2004). We believe that the cross-correlation method used in the present study is a better indicator of temporal resolving power in studies using recordings
AVMA Panel on Euthanasia, American Veterinary Medical Association 2000 Report of the AVMA Panel on Euthanasia, J Am Vet Med Assoc 2001; 218: 669–696. Bartol SM, Musick JA. Sensory biology of sea turtles. In: Lutz PL, Musick JA, Wyneken J, editors. The Biology of sea turtles, vol. II. Boca Raton, FL: CRC Press; 2003. p. 79–102. Bjorndal KA. Foraging ecology and nutrition of sea turtles. In: Lutz PL, Musick JA, editors. The biology of sea turtles. Boca Raton, FL: CRC Press; 1997. p. 199–231. Bolten AB. Active swimmers—passive drifters: the oceanic juvenile stage of loggerheads in the Atlantic system. In: Bolten AB, Witherington BE, editors. Loggerhead sea turtles. Washington, DC: Smithsonian Books; 2003. p. 63–78. Constantino MA, Salmon M. Role of chemical and visual cues in food recognition by leatherback posthatchlings (Dermochelys coriacea L.). Zoology 2003;106:173–81. Crognale MA, Eckert SA, Levenson DH, Harms CA. Leatherback sea turtle Dermochelys coriacea visual capacities and potential reduction of by-catch by pelagic longline fisheries. Endangered Species Research 2008;5:249–56. Eckert S, Levenson DH, Cognale MA. The sensory biology of sea turtles: what can they see, and how can this help them avoid fishing gear? In: Swimmer Y, Brill R, editors. Sea turtle and pelagic fish sensory biology: developing techniques to reduce sea turtle bycatch in longline fisheries. Washington, DC: NOAA Technical Memorandum, NOAA-TM-NMFS-PIFSC-7, U.S. Dept. Commerce; 2006. p. 8–16. Eckert SA, Eckert KL. Diving and foraging behavior of leatherback sea turtles (Dermochelys coriacea). Can J Zool 1989;67:2834–40. Eckert SA, Nellis DW, Eckert KL, Kooyman GL. Diving patterns of two leatherback sea turtles (Dermochelys coriacea) during internesting intervals at Sandy Point, St. Croix, U.S. Virgin Island. Herpetologica 1986;42:381–8. Fritsches KA, Brill RW, Warrant EJ. Warm eyes provide superior vision in swordfishes. Curr Biol 2005;15:55–8. Fritsches KA, Warrant EJ. Differences in the visual capabilities of sea turtles and blue water fishes—implications for bycatch reduction. In: Swimmer Y, Brill R, editors. Sea turtle and pelagic fish sensory biology: developing techniques to reduce sea turtle bycatch in longline fisheries. Washington, DC: NOAA Technical Memorandum, NOAA-TM-NMFS-PIFSC-7, U.S. Dept. Commerce; 2006. p. 1–7. Granda AM. Eyes and their sensitivity to light of differing wavelengths. In: Harless M, Morlock H, editors. Turtles, perspectives and research. New York: John Wiley & Sons; 1979. p. 247–66. Grant GS, Ferrell D. Leatherback turtle, Dermochelys coriacea (Reptilia: Dermochelidae): notes on nearshore feeding behavior and association with cobia. Brimleyana 1993;19:77–81. Hecht S, Shlaer S. Intermittent stimulation by light. V. Relation between intensity and critical frequency for different parts of the spectrum. J Gen Physiol 1936;19:965–70. Heithaus MR, Frid A, Dill LM. Shark-inflicted injury frequencies, escape ability, and habitat use of green and loggerhead turtles. Marine Biol 2002;140:229–36.
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283 Horch KW, Gocke JP, Salmon M, Forward RB. Visual spectral sensitivity of hatchling loggerhead (Caretta caretta L.) and leatherback (Dermochelys coriacea L.) sea turtles, as determined by single-flash electroretinography. Marine Freshwater Behav Physiol 2008;41:107–19. Jacobs GH, Neitz J, Krogh K. Electroretinogram flicker photometry and its applications. J Opt Soc Am A 1996;13:641–8. James MC, Ottensmeyer CA, Eckert SA, Myers RA. Changes in diel diving patterns accompany shifts between northern foraging and southward migration in leatherback turtles. Can J Zool 2006;84:754–65. Levenson DH, Eckert SA, Crognale MA, Deegan JFI, Jacobs GH. Photopic spectral sensitivity of green and loggerhead sea turtles. Copeia 2004;2004:908–14. Levine MW. Cross-correlation between neurons: a source of information about the nervous system. Biosystems 1998;48:139–46. Mäthger LM, Litherland L, Fritsches KA. An anatomical study of the visual capabilities of the green turtle, Chelonia mydas. Copeia 2007;2007:169–79. Perlman I, Normann RA, Chandler JP, Lipetz LE. Effects of calcium ions on L-type horizontal cells in the isolated turtle retina. Vis Neurosci 1990;4:53–62. Salmon M. Artificial night lighting and sea turtles. The Biologist 2003;50:163–8. Salmon M, Jones TT, Horch KW. Ontogeny of diving and feeding behavior in juvenile seaturtles: leatherback seaturtles (Dermochelys coriacea L.) and green seaturtles (Chelonia mydas L.) in the Florida current. J Herpetol 2004;38:36–43.
283
Southwood A, Fritsches K, Brill R, Swimmer Y. Sound, chemical, and light detection in sea turtles and pelagic fishes: sensory-based approaches to bycatch reduction in longline fisheries. Endangered Species Res 2008;5, preprint doi:10.3354/esr00097. Witherington BE. Behavioral responses of nesting sea turtles to artificial lighting. Herpetologica 1992;48:31–9. Witherington BE. Ecology of neonate loggerhead turtles inhabiting lines of down welling near a Gulf Stream front. Marine Biol 2002;140:843–53. Witherington BE. Orientation of hatchling loggerhead turtles at sea off artificially lighted and dark beaches. J Exp Marine Biol Ecol 1991;149:1–11. Witherington BE. The problem of photopollution for sea turtles and other nocturnal animals. In: Clemmons JR, Buchholz R, editors. Behavioral approaches to conservation in the wild. Cambridge, UK: Cambridge University Press; 1997. p. 303–28. Witherington BE, Martin RE. Understanding, assessing, and resolving light-pollution problems on sea turtle nesting beaches. Florida Marine Research Institute Technical Report TR-2; 1996. Wyneken J, Salmon M. Frenzy and post frenzy swimming activity in loggerhead, green, and leatherback hatchling sea turtles. Copeia 1992;1992: 478–84.
Contents lists available at ScienceDirect
Journal of Neuroscience Methods journal homepage: www.elsevier.com/locate/jneumeth
Frequency response characteristics of isolated retinas from hatchling leatherback (Dermochelys coriacea L.) and loggerhead (Caretta caretta L.) sea turtles Kenneth Horch a,∗ , Michael Salmon b a b
Department of Bioengineering, University of Utah, 50 S. Central Campus Dr., Salt Lake City, UT 84112, USA Department of Biological Sciences, Florida Atlantic University, Box 3091, 777 Glades Road, Boca Raton, FL 33431-0991, USA
a r t i c l e
i n f o
Article history: Received 20 August 2008 Received in revised form 12 December 2008 Accepted 12 December 2008 Keywords: Vision Electroretinogram Marine turtle Leatherback Loggerhead Frequency sensitivity
a b s t r a c t Electroretinographic recordings were made from hatchling loggerhead and leatherback sea turtle eyecup preparations during presentation of sinusoidally modulated lights of different frequencies, mean intensities and colors. Cross-correlation analysis was performed to determine the extent to which the responses followed the intensity modulated light sources. For both species mean light intensity had no significant effect on the frequency modulated responses over a 1.5 log unit range of intensities. Both species showed the best following to blue light and the poorest tracking to red light. Leatherback retinas did not follow frequencies above 10 Hz, while loggerhead responses extended out to 15 Hz. These visual low pass filter characteristics are consistent with attributes of the visual ecology of each species, as well as with the latencies and slow rise times exhibited by these retinas to brief flashes of light. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Vision plays an important role in many aspects of marine turtle behavior such as orientation (Salmon, 2003; Witherington, 1991, 1997; Witherington and Martin, 1996), habitat selection (Witherington, 1992), as well as detection of prey (Constantino and Salmon, 2003; Salmon et al., 2004) and predators (Heithaus et al., 2002). Even so, our understanding of relationships between visual physiology, ecology and behavior among marine turtles is in its early stages (Bartol and Musick, 2003). It is known that sea turtles possess multiple visual pigments (Granda, 1979) and potentially can discriminate between wavelengths that extend from the near UV to red, although this has yet to be demonstrated behaviorally (Crognale et al., 2008; Horch et al., 2008; Levenson et al., 2004; Mäthger et al., 2007). An important characteristic of visual systems is their ability to resolve rapid changes in light intensity (their visual temporal bandwidth: informally their flicker fusion frequency). Fritsches (Fritsches and Warrant, 2006; Southwood et al., 2008) has reported a surprisingly high upper frequency limit near 60 Hz for adult green turtle retinas. This is at variance with reports of others who found bandwidths at or below 20 Hz in green turtle retinas, and even lower cutoff frequencies in leatherbacks (Dermochelys coriacea L.) and loggerheads (Caretta caretta L.), during flicker photometry mea-
∗ Corresponding author. Tel.: +1 801 585 1981; fax: +1 801 585 5361. E-mail addresses: [email protected] (K. Horch), [email protected] (M. Salmon). 0165-0270/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2008.12.017
surements (Crognale et al., 2008; Eckert et al., 2006; Levenson et al., 2004). Their results are consistent with our study of retinal spectral sensitivity in hatchling leatherback and loggerhead sea turtles. We found response latencies to brief light flashes on the order of 50–100 ms, and rise times to peak of the photoreceptor (a-wave) responses on the order of 100–300 ms (Horch et al., 2008). Although the flicker photometry and spectral sensitivity work did not directly address the issue of temporal bandwidth, these results do differ from what would be expected if a 60 Hz frequency following limit were correct. We believe that a primary reason for this discrepancy is methodological and, in particular, that examining “flicker fusion frequency” using Fourier transform and related techniques is not as reliable as one would like. Our intention here is to present a more mathematically robust approach. Leatherbacks and loggerheads are of interest not only because they are endangered species for which vision plays an important role in survival, especially early during ontogeny, but because differences in their visual ecology and behavior suggest their eyes could exhibit differences in temporal resolving power. Leatherbacks as hatchlings are active primarily during the day (Wyneken and Salmon, 1992). As adults, they feed during the day and at night (Eckert and Eckert, 1989; Eckert et al., 1986; James et al., 2006) at the surface where illumination is bright (Grant and Ferrell, 1993) or by diving to depths >1000 m where down welling light is absent (Eckert and Eckert, 1989; Eckert et al., 1986). Leatherbacks during all stages of ontogeny are feeding specialists consuming slow moving gelatinous prey (jellyfishes, medusae, ctenophores, and salps) (Bjorndal, 1997).
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
Loggerheads as hatchlings are also day-active (Wyneken and Salmon, 1992). Juveniles, during their pelagic stage of development, feed primarily upon stationary prey gleaned from flotsam floating at the surface (Witherington, 2002) or found at shallow depths in the water column (Bolten, 2003). Older turtles most often return to coastal waters where they primarily consume sessile benthic prey (bivalves, soft corals, anemones), gelatinous zooplankton in the water column, and prey capable of more rapid movement (horseshoe and hermit crabs, mollusks, seahorses, portunid crabs) (Bjorndal, 1997). These ecological considerations led us to speculate that loggerheads might have somewhat “faster” eyes than leatherbacks. Additionally, most information about sea turtle vision stems from work on green turtles, with very little information available about either leatherback or loggerhead turtles at young ages (Horch et al., 2008). It is of interest to know the extent to which observations about visual physiology from one species or life stage of sea turtles can be extrapolated to other species and life stages. Both species have more than one retinal pathway (Horch et al., 2008). The existence of separate pathways might result in colordependent differences in visual temporal resolution, a concept that to the best of our knowledge has not been previously addressed in marine turtles. The claim that color effects are absent in human perceptual flicker fusion frequency is based upon the negative results obtained by Hecht and Shlaer (1936), who used themselves as subjects. The present study was designed to examine the ability of hatchling leatherback and loggerhead retinas to track changes in light intensity as a function of stimulus tempo and color, using a standard system analysis approach. This involved a mathematically well-defined stimulus (sinusoidally modulated light intensity) and a mathematically appropriate direct measure of the relationship between the stimulus and retinal response (cross-correlation analysis). Correlation analysis is a widely used and accepted method for examining information processing in the nervous system: it provides a direct measure of the relationship between two events (in this case, the stimulus and the recorded electrical response of the retina) in the form of the correlation coefficient (Levine, 1998). Additionally, it provides a direct measure of the temporal relationship between the two events (in this case, the lag or latency between the stimulus and the response). The latter is important not only for confirming that the recorded responses have latencies similar to those predicted by more direct measures of retinal response latencies (Horch et al., 2008), but also for identifying and, in appropriate circumstances, eliminating stimulus artifacts from the recordings when performing the data analysis. 2. Materials and methods Because both species are endangered worldwide, our permit required us to use no more than 6 “stragglers” of each species as subjects. Stragglers are hatchlings remaining inside the nest 72 h after the majority of the turtles have departed and are migrating offshore; turtles left behind usually do not survive, even with human intervention. All were collected during daylight hours between July and September, 2005 and 2006, from nests in Palm Beach and Broward Counties, Florida, USA. The turtles were stored in a covered Styrofoam cooler in a dark, air-conditioned laboratory. A shallow layer of moist sand or a wet paper towel was placed inside the cooler to prevent desiccation. They were sacrificed for experiments within 12 h after capture. Although the hatchlings were incapable of leaving the nest, their eyes seemed fully developed and they showed consistent responses to the visual stimuli we presented. Initial experiments were performed on two loggerheads to hone our techniques. Final experiments were done on the remaining animals permitted by our protocol (6 leatherbacks and 4 loggerheads).
277
2.1. Retinal preparations Hatchlings were anesthetized in an ice water bath for ∼20 min, or until they reached a state of torpor, decapitated at the cervical junction, and pithed according to accepted protocols (AVMA Panel on Euthanasia. American Veterinary Medical Association, 2001). The cornea, lens and vitreous humor of one eye were removed under indirect, dim illumination of a dissecting microscope (10–25×), leaving a retinal eyecup, which was filled with perfluoro FC-77 ¯ Organics) to prevent desiccation and promote oxygen diffu(Acros sion to the retina. The preparation was placed inside a Faraday cage and a Pt ground-wire was placed inside the skull. The preparation was enclosed in a clear plastic “tent”; fully humidified air mixed with 95% oxygen continuously flowed into the tent. The preparation was dark-adapted for 20 min before light stimuli were presented (Perlman et al., 1990). See Horch et al., 2008 for a more detailed description. 2.2. Recording and stimulation equipment A chloridized silver wire loop was placed in contact with the periphery of the retina. Preliminary tests confirmed that there were no responses to visual stimuli arising from the electrode itself. Signals from the retina were fed to a modified WPI DAM-50 preamplifier (gain: 1000×; bandpass: 1–1 kHz). The preamplified electroretinographic signals were monitored on a storage oscilloscope and sent to one channel of a 16-bit A/D card that allowed acquisition of the data on a laptop computer (1000 samples/s). After each stimulus sequence, the response was displayed on the computer screen. Full amplitude, sinusoidal waveforms at the desired frequency were generated by the computer and offset appropriately to provide a roughly 2.4 log unit intensity modulation. The waveform was fed to a step attenuator (attenuation range 0–2.5 log units in 0.5 log unit steps) through a digital-to-analog converter. The attenuated signal was used to control a constant current source that drove the selected LED stimulus light source. Four LEDs were used (all manufactured by Purdy Electronics Corp., Sunnyvale, CA): AND190HBA blue (430 nm peak), AND412HG blue-green (505 nm peak), AND185HOP orange (620 nm peak), and AND190HW white (significant output from 450 to 650 nm). LED spectral outputs are shown in Fig. 1. The LEDs were mounted on a carrier attached to a wafer switch so that changing from one light source to another simply required rotating the switch. This action simultaneously placed the appropriate LED at the same position relative to the eyecup and directed the driving current to it. An Ocean Optics S-2000 spectrometer was used to confirm wavelength specifications and that, due to the quantum nature of the light emission process, the color remained constant over the range of intensities used in the experiments. A UDT radiometer (model S351A, uniform response between 400 and 700 nm) was used to measure light intensity at the retina in W/cm2 . Light intensities derived from these measurements were used in all data analysis. 2.3. Stimulation and recording procedures A stimulus sequence consisted of providing a steady illumination at the mean intensity of the modulated light for 1–3 s. The duration of this steady pre-stimulus depended on the preparation and was chosen to be long enough to allow the transient response to the onset of the light to decay away. The light intensity was then varied sinusoidally (modulation depth approximately 2.4 log units) for 2 s. The signal recorded from the retina during this 2 s period was recorded on the computer, but only the last 1 s was used in the data analysis because the non-linear response properties of the
278
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
Fig. 1. Emission spectra for the LED light sources used in this study. Plots taken from manufacturer’s data sheets.
eye produced a transient response that could last up to 1 s once the modulation began. The light intensities were chosen to be bright enough to elicit photopic (i.e., cone) responses (Horch et al., 2008), but not bright enough to saturate the visual responses. Stimulus color and intensity modulation rate were presented randomly to the different preparations, but for a given color and rate, we always started with the lowest intensity and subsequently increased intensity in 0.5 log unit steps from there. Each stimulus combination was presented three times, except in cases where a visible artifact was present. In those instances additional repetitions were used to provide three “clean” records. 2.4. Data analysis Cross-correlation analysis provided a quantitative measure of the ability of the eyes to follow changes in light intensity, including accommodation of non-linearities in their stimulus–response properties, compensation for differences in recorded response amplitudes, and maintenance of temporal information about response latency. The square of the peak correlation coefficient (r2 ) is a measure of the fraction of the variability in the response that is due to variability in the stimulus (i.e., how well the system followed the stimulus). This approach is self-scaling as it measures the ability of the retina to follow intensity changes in the stimulus, largely independent of response amplitude. To find the peak correlation coefficient, the analysis was performed with different latencies (lags) between the stimulus waveform and the response waveform. From our data on response latencies to flashes of lights in these eyes (Horch et al., 2008), we expected response latencies or lags to be less than 200 ms, and so
looked at shifts between 0 and 200 ms. A very short latency would indicate a stimulus artifact, and a lag much greater than 100 ms would indicate something other than a neural response from the retina. Because the photoreceptor response to light in this preparation is negative and has a shorter time constant than the later response components (Horch et al., 2008), we expected to see a peak, negative correlation coefficient at frequencies of 10 Hz and above appearing somewhere in the vicinity of a 100 ms lag. Cross-correlation analysis was performed on the last 1 s of data using custom Matlab® scripts. The stimulus and response waveforms were first normalized to given autocorrelation values of 1. The peak negative cross-correlation coefficient was found and its square was used as the metric for how well the response followed the stimulus. Although not evident on visual observation of the responses, examination of the cross-correlation behavior revealed that there was a stimulus artifact present in the data. The properties of the artifact indicated that it was likely due to cross-talk between the stimulus waveform sent out the D/A converter and the incoming signal digitized by the A/D converter. Since the former was constant in amplitude and exhibited no lag, we were able to determine its characteristics for each recording session and generate its inverse mathematically. This inverse was added to the original recorded signal, thus canceling out the artifact, before performing the final cross-correlation analysis for the data presented here. 3. Results The top two panels in Fig. 2 show recordings and their crosscorrelograms made in response to three intensities of blue light
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
279
Fig. 2. Example of a data acquisition and analysis sequence. The top left panel shows responses from a leatherback hatchling eyecup to blue light sinusoidally modulated at 5 Hz around three different mean intensities: the lighter the trace the brighter the light. The stimulus (normalized) is presented as the trace at the bottom of the panel. The top right panel shows the cross-correlations (r values) from these recordings (calculated at different lags (latencies) between stimulus and response, uncorrected for stimulus artifact). The middle row shows the same for 7 Hz stimulation. Note that in both cases there is a decrease in response latency, indicated by a leftward shift of the peak (negative) r value, with increasing light intensity, but the peak r values are independent of response amplitude. The bottom panels are plot of the squares of the maximum cross-correlation values derived from these three light intensities over all frequencies tested with the blue LED in this animal, without (left) and with (right) artifact correction. Shading differentiates values above from those below 0.5.
modulated at 5 Hz presented to one of the leatherback eyecup preparations. The second row shows similar data for light modulated at 7 Hz. Response amplitudes decreased with increasing mean light intensity. This was probably due, either alone or in combination, to the logarithmic nature of the visual response to light intensity; to changes in light adaptation of the preparation with changes in mean level of illumination; or to increasing B-wave contribution to the response, which is opposite in sign to the photoreceptor response (Horch et al., 2008). However, the correlation values were largely the same at a given modulation rate—indicating that the eye’s ability to follow the changes in light intensity was unaffected by light intensity over this range of intensities. As seen in our earlier study, response latencies decreased with increasing stimulus intensity (Horch et al., 2008). The bottom row of the figure shows plots of the squares of the peak correlation values (without and with correction for stimulus artifact) for all frequencies tested with blue light at these three intensities in this preparation.
The left column in Fig. 3 shows three-dimensional plots of averaged response profiles for all 6 leatherbacks for each of the four colors used. On the right these data are presented in twodimensional format with standard errors. There is no consistent or statistically significant difference between the four curves for a given color. Fig. 4 shows corresponding data for four loggerhead hatchlings. As with the leatherbacks, the ability to follow light modulation did not depend on stimulus intensity. Data for each of the four stimulus intensities for each color were combined and the results are shown in Fig. 5. In both species, the frequency following curve for red light lies below the other curves (except at high rates where all the curves are approaching an asymptote and there is, essentially, no following of the stimulus for any color), an effect that is more pronounced in leatherbacks than in loggerheads. The former shows a general trend for the retinas to follow blue light better than green light better than red. However, for the loggerheads, the curves for blue and green light overlap.
280
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
Fig. 3. Average (n = 6) of squares of peak cross-correlation values as a function of light color, modulation frequency and mean intensity for retinas from leatherback hatchlings. Curves on the right, extracted from the three-dimensional plots on the left, show standard errors. The lighter the trace, the brighter the light. Data points offset for clarity.
4. Discussion Neither species appears capable of following light flicker at frequencies much above 15 Hz. Hatchling leatherback retinas respond to light modulation at 10 Hz about as well as loggerheads do at 15 Hz, supporting our suggestion that loggerhead eyes, while slow, are “faster” than leatherback eyes.
Evidence for color-dependent differences in visual temporal resolution is less obvious, but suggested by the consistently lower correlation coefficients for red light modulation than those for the other colors. These results suggest that the retinal components responding to long wavelengths in both species may be slower than those responding to shorter wavelengths. We speculate that this effect is unlikely to arise as a consequence of differences in visual
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
281
Fig. 4. Data from loggerheads (n = 4). Format as in Fig. 3.
sensitivity to red versus blue and green lights for the following reasons. First, if sensitivity was a factor, the r2 values would depend on stimulus intensity (the level of effective retinal stimulation). However, these values were independent of stimulus intensity at the intensities used here. Second, the quantal content of the bright red light was 1.2 log units higher than that of the dim blue light, a bigger difference than the difference in retinal sensitivities to the red versus the blue light (−0.9 log units for the leatherbacks, −0.2 log
units for the loggerheads) (Horch et al., 2008). Finally, the ranges of response amplitudes to red and blue lights at 10 Hz were indistinguishable from each other within a species, meaning that the measurements for different colors were effectively made at comparable response amplitudes. Such a temporal resolving power difference may occur because the longer wavelength pathways converge more to compensate for lower sensitivity to red light, at the expense of temporal coher-
282
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283
of whole or intact retinal responses to changes in light intensity. In the final analysis, however, behavioral tests of the animals’ ability to perceive modulations in light intensity are needed demonstrate the extent to which animals are actually aware of and can respond to flickering lights. In conclusion, we demonstrate in this study that combining the use of sinusoidally modulated light intensity with cross-correlation analysis is a valuable method for assessing temporal resolving power in visual systems. This procedure allowed us to detect a stimulus artifact that otherwise would have provided skewed results, decreasing apparent correlation when the artifact was out of phase with the visual responses and increasing apparent correlation when the artifact was in phase with any responses or when a real visual response was absent (Fig. 2, bottom row). This method also compensates for differences in recorded signal amplitudes, provided that the responses are sufficiently large that the inherent signal averaging in the procedure can pull them out of the noise. Finally, it also provides a measure with direct physical significance: r2 is an indication of the extent to which any recorded signals are actually following the stimulus. Acknowledgments Permits to carry out this project were obtained from the U.S. Fish and Wildlife Service (TE087169-0), the Florida Fish and Wildlife Conservation Commission (TP # 173), and the Florida Atlantic University Institutional Animal Care Committee (A02-22). Support was provided by the NOAA Fisheries Sensory Biology Working Group. We thank A. McDermott for assistance with the experiments and T. Frank for comments that helped improve the manuscript. References
Fig. 5. Combined data from Figs. 3 and 4, with standard errors, for leatherback (top) and loggerhead (bottom) hatchlings.
ence in their inputs. This adaptation would be consistent with the environment at depths where these animals feed during dives, and where the visible spectrum shifts towards the blue and UV wavelengths. Note that the effects illustrated in Fig. 2 lead to two caveats. The first, and most obvious, is that the amplitude of the response, measured from whole retina preparations, is not a good indicator of frequency following by the eye. For example, the use of power spectra derived from a Fourier analysis of retinal recordings as a criterion for determining flicker fusion frequency (Fritsches et al., 2005), in addition to its theoretical and proper implementation difficulties, is basically an amplitude measure and should be employed with caution. The second, less obvious, point is that using responses to amplitude modulated light sources to estimate spectral sensitivity (Jacobs et al., 1996) may be subject to error if one does not consider that the different color pathways may show different temporal bandwidths. This is supported by the work of Crognale et al. (2008). Fig. 3 of that paper demonstrates that leatherbacks exhibit a dramatic difference in spectral sensitivity to long wavelengths depending upon the stimulation rate (8 Hz vs. 12 Hz). In addition, these two spectral sensitivities are different from one derived using an “ERG criterion” presented (Fig. 2) in the same paper. While our results differ from the significantly higher flicker fusion frequencies reported by Fritsches (Fritsches and Warrant, 2006; Southwood et al., 2008), they are consistent with related findings in green, leatherback and loggerhead turtles (Crognale et al., 2008; Eckert et al., 2006; Levenson et al., 2004). We believe that the cross-correlation method used in the present study is a better indicator of temporal resolving power in studies using recordings
AVMA Panel on Euthanasia, American Veterinary Medical Association 2000 Report of the AVMA Panel on Euthanasia, J Am Vet Med Assoc 2001; 218: 669–696. Bartol SM, Musick JA. Sensory biology of sea turtles. In: Lutz PL, Musick JA, Wyneken J, editors. The Biology of sea turtles, vol. II. Boca Raton, FL: CRC Press; 2003. p. 79–102. Bjorndal KA. Foraging ecology and nutrition of sea turtles. In: Lutz PL, Musick JA, editors. The biology of sea turtles. Boca Raton, FL: CRC Press; 1997. p. 199–231. Bolten AB. Active swimmers—passive drifters: the oceanic juvenile stage of loggerheads in the Atlantic system. In: Bolten AB, Witherington BE, editors. Loggerhead sea turtles. Washington, DC: Smithsonian Books; 2003. p. 63–78. Constantino MA, Salmon M. Role of chemical and visual cues in food recognition by leatherback posthatchlings (Dermochelys coriacea L.). Zoology 2003;106:173–81. Crognale MA, Eckert SA, Levenson DH, Harms CA. Leatherback sea turtle Dermochelys coriacea visual capacities and potential reduction of by-catch by pelagic longline fisheries. Endangered Species Research 2008;5:249–56. Eckert S, Levenson DH, Cognale MA. The sensory biology of sea turtles: what can they see, and how can this help them avoid fishing gear? In: Swimmer Y, Brill R, editors. Sea turtle and pelagic fish sensory biology: developing techniques to reduce sea turtle bycatch in longline fisheries. Washington, DC: NOAA Technical Memorandum, NOAA-TM-NMFS-PIFSC-7, U.S. Dept. Commerce; 2006. p. 8–16. Eckert SA, Eckert KL. Diving and foraging behavior of leatherback sea turtles (Dermochelys coriacea). Can J Zool 1989;67:2834–40. Eckert SA, Nellis DW, Eckert KL, Kooyman GL. Diving patterns of two leatherback sea turtles (Dermochelys coriacea) during internesting intervals at Sandy Point, St. Croix, U.S. Virgin Island. Herpetologica 1986;42:381–8. Fritsches KA, Brill RW, Warrant EJ. Warm eyes provide superior vision in swordfishes. Curr Biol 2005;15:55–8. Fritsches KA, Warrant EJ. Differences in the visual capabilities of sea turtles and blue water fishes—implications for bycatch reduction. In: Swimmer Y, Brill R, editors. Sea turtle and pelagic fish sensory biology: developing techniques to reduce sea turtle bycatch in longline fisheries. Washington, DC: NOAA Technical Memorandum, NOAA-TM-NMFS-PIFSC-7, U.S. Dept. Commerce; 2006. p. 1–7. Granda AM. Eyes and their sensitivity to light of differing wavelengths. In: Harless M, Morlock H, editors. Turtles, perspectives and research. New York: John Wiley & Sons; 1979. p. 247–66. Grant GS, Ferrell D. Leatherback turtle, Dermochelys coriacea (Reptilia: Dermochelidae): notes on nearshore feeding behavior and association with cobia. Brimleyana 1993;19:77–81. Hecht S, Shlaer S. Intermittent stimulation by light. V. Relation between intensity and critical frequency for different parts of the spectrum. J Gen Physiol 1936;19:965–70. Heithaus MR, Frid A, Dill LM. Shark-inflicted injury frequencies, escape ability, and habitat use of green and loggerhead turtles. Marine Biol 2002;140:229–36.
K. Horch, M. Salmon / Journal of Neuroscience Methods 178 (2009) 276–283 Horch KW, Gocke JP, Salmon M, Forward RB. Visual spectral sensitivity of hatchling loggerhead (Caretta caretta L.) and leatherback (Dermochelys coriacea L.) sea turtles, as determined by single-flash electroretinography. Marine Freshwater Behav Physiol 2008;41:107–19. Jacobs GH, Neitz J, Krogh K. Electroretinogram flicker photometry and its applications. J Opt Soc Am A 1996;13:641–8. James MC, Ottensmeyer CA, Eckert SA, Myers RA. Changes in diel diving patterns accompany shifts between northern foraging and southward migration in leatherback turtles. Can J Zool 2006;84:754–65. Levenson DH, Eckert SA, Crognale MA, Deegan JFI, Jacobs GH. Photopic spectral sensitivity of green and loggerhead sea turtles. Copeia 2004;2004:908–14. Levine MW. Cross-correlation between neurons: a source of information about the nervous system. Biosystems 1998;48:139–46. Mäthger LM, Litherland L, Fritsches KA. An anatomical study of the visual capabilities of the green turtle, Chelonia mydas. Copeia 2007;2007:169–79. Perlman I, Normann RA, Chandler JP, Lipetz LE. Effects of calcium ions on L-type horizontal cells in the isolated turtle retina. Vis Neurosci 1990;4:53–62. Salmon M. Artificial night lighting and sea turtles. The Biologist 2003;50:163–8. Salmon M, Jones TT, Horch KW. Ontogeny of diving and feeding behavior in juvenile seaturtles: leatherback seaturtles (Dermochelys coriacea L.) and green seaturtles (Chelonia mydas L.) in the Florida current. J Herpetol 2004;38:36–43.
283
Southwood A, Fritsches K, Brill R, Swimmer Y. Sound, chemical, and light detection in sea turtles and pelagic fishes: sensory-based approaches to bycatch reduction in longline fisheries. Endangered Species Res 2008;5, preprint doi:10.3354/esr00097. Witherington BE. Behavioral responses of nesting sea turtles to artificial lighting. Herpetologica 1992;48:31–9. Witherington BE. Ecology of neonate loggerhead turtles inhabiting lines of down welling near a Gulf Stream front. Marine Biol 2002;140:843–53. Witherington BE. Orientation of hatchling loggerhead turtles at sea off artificially lighted and dark beaches. J Exp Marine Biol Ecol 1991;149:1–11. Witherington BE. The problem of photopollution for sea turtles and other nocturnal animals. In: Clemmons JR, Buchholz R, editors. Behavioral approaches to conservation in the wild. Cambridge, UK: Cambridge University Press; 1997. p. 303–28. Witherington BE, Martin RE. Understanding, assessing, and resolving light-pollution problems on sea turtle nesting beaches. Florida Marine Research Institute Technical Report TR-2; 1996. Wyneken J, Salmon M. Frenzy and post frenzy swimming activity in loggerhead, green, and leatherback hatchling sea turtles. Copeia 1992;1992: 478–84.