Copeia, 2004(4), pp. 908–914 Photopic Spectral Sensitivity of Green and Loggerhead Sea Turtles D. H. LEVENSON,S.A.ECKERT,M.A.CROGNALE,J.F.DEEGAN II, AND G. H. JACOBS Flicker electroretinography (ERG) was used to examine the in situ photopic (cone- photoreceptor based) spectral sensitivities of Green and Loggerhead Sea Turtles. Both species were responsive to wavelengths from 440–700 nm, and both had peak sensitivity in the long wavelength portion of the spectrum (ϳ580 nm). For Logger- head Sea Turtles, no measurable responses were obtained below about 440 nm, whereas reliable signals were seen for Green Sea Turtles at wavelengths down to 400 nm. Both species exhibited significant declines in sensitivity below 500 nm. The overall shapes of the spectral sensitivity functions were similar for the two species. These results support previous findings that sea turtles have well-developed phot- opic visual systems. The characteristics of these spectral sensitivity functions indicate that both species possess multiple cone photopigment types, and these, in conjunc- tion with the presence of colored oil droplets, strongly imply a capacity for color discrimination. Comparative evaluation suggests that these turtles have modified their visual pigments from those of their terrestrial relatives to better suit the am- bient conditions present in the shallow water, submarine environments that they typically inhabit. ODERN sea turtles diverged from their wavelength sensitive visual pigments of marine M terrestrial and freshwater relatives rough- turtles toward shorter wavelengths. MSP also re- ly 150 million years ago (Pritchard, 1997). They veals that the light-absorbing oil droplets pre- live an almost exclusively marine existence, re- sent in marine turtle cone photoreceptors are turning to the ocean within a few hours of different from those of terrestrial and freshwa- hatching and subsequently rarely returning to ter turtles (Granda and Haden, 1970; Granda land. Sea turtles are currently distributed world- and Dvorak, 1977). wide throughout the tropical and subtropical Although there is evidence that the visual oceans, with some species occasionally ranging sense is important to sea turtles throughout all into higher, more temperate latitudes. Sea tur- stages of their life cycle (Ehrenfeld and Carr, tles occupy a variety of ecologic niches. For in- 1967; Kingsmill and Mrosovsky, 1982; Lohmann stance, adult Green Sea Turtles (Chelonia mydas) et al., 1990), relatively little is known about their are generally found near shore feeding on plant in situ visual sensitivities; most previous work and algal matter, whereas others, such as adult has been directed at the individual photorecep- Loggerhead Sea Turtles (Caretta caretta) may tors. Because several conservation issues cur- stay farther offshore, diving deeper to forage on rently facing marine turtles can be related, at benthic invertebrates and fish (Bjorndal, 1997). least in part, to the effects of anthropogenic Anatomical examinations reveal that, in con- light sources, there is a particular need for a trast to terrestrial and freshwater turtles, the better understanding of the visual abilities of liv- eyes of sea turtles have round lenses, much like ing turtles (e.g., Witherington, 1991; Witzell, those found in fish, to compensate for the loss 1999). As a step in this direction, we have used of corneal refraction under water (Walls, 1942; a noninvasive technique, rapid-flicker electro- Ehrenfeld and Koch, 1967; Northmore and retinography (ERG), to investigate the photopic Granda, 1991). The spectral sensitivity of ma- (cone-photoreceptor based) visual sensitivities rine turtles is also different from that of their of two marine turtle species, Green and Log- land-based relatives (Liebman and Granda, gerhead Sea Turtles. This type of ERG is partic- 1971; Granda and Dvorak, 1977). Microspectro- ularly relevant to behavioral/conservation is- photometric (MSP) evaluations of Green Sea sues, because it reflects the actual sensitivity of Turtle photoreceptors show that 11-cis retinal is the eyes of the animal, including the combined used as the light sensitive chromophore com- effects of the cone visual pigments, oil droplets, ponent of sea turtle visual pigments; land based and other sources of intraocular scattering and turtles typically use 3,4 di-dehydroretinal (Lieb- absorption (e.g. cornea, lens). Rapid-flicker man and Granda, 1971; Loew and Govardovski, ERG was chosen over the single flash ERG tech- 2001). This substitution has the effect of sub- nique previously used with Green Sea Turtles stantially shifting the sensitivity of the longer (Granda and O’Shea, 1972), because it is better ᭧ 2004 by the American Society of Ichthyologists and Herpetologists LEVENSON ET AL.—SEA TURTLE VISUAL PIGMENTS 909 suited to isolating cone photoreceptor respons- ments, the eye was stimulated with a train of es and thereby potentially providing a more ac- light pulses originating from a high-intensity curate representation of in situ photopic sensi- (50-W tungsten-halide lamp) grating mono- tivity. Furthermore, by conducting parallel ex- chromator having a half bandpass of 15 nm. periments with two species, we reasoned that, The light was imaged onto the retina in Max- because the individual cone photopigments of wellian view in the form of a spot subtending the Green Sea Turtle have been previously mea- 59 deg. The test light was temporally modulated sured, we could use the ERG measurements on with electromechanical shutters (Vincent Asso- Green Sea Turtles to draw more precise infer- ciates, Rochester, NY) as a square-wave pulse ences about the photoreceptor complements of having a 25% duty cycle so as to achieve any Loggerhead Sea Turtles. desired flicker rate. The fundamental frequency component of the ERG response was extracted MATERIALS AND METHODS by filtering and this signal was averaged over a total of 50 presentations. Spectral sensitivity Live adult (all greater than 30 years old) functions were measured using two techniques. Green and Loggerhead Sea turtles were exam- In the first case, a flicker photometric proce- ined at SeaWorld, San Diego, California. Re- dure was employed in which the responses to cordings were obtained from four Green Sea the test light and those given to an interleaved Turtles and six Loggerhead Sea Turtles. It was reference light (achromatic, 14 log photons/ not possible to determine the gender of some sec/sr) that illuminated the same region of the of the turtles, but at least one male and one retina and flickered at the same temporal fre- female were examined from each species. These quency were compared. Over successive presen- animals were maintained in captivity in seawater tations, the intensity of the monochromatic test exhibits before and after recording. All animal light was adjusted by changing the position of a husbandry and experimental procedures were 3.0-log unit neutral-density wedge until the light conducted according to protocols approved by produced a response equal to that given to the the SeaWorld Institutional Animal Care and Use fixed reference light. Repetition of this proce- Committee (IACUC) and were overseen by the dure for a range of different test wavelengths veterinary staff at SeaWorld. can be used to define a spectral sensitivity func- Photopic (cone-photoreceptor based) spec- tion ( Jacobs and Neitz, 1987). In a second set tral sensitivity was evaluated for each turtle in of measurements, spectral sensitivity functions vivo using flicker electroretinography (ERG). were similarly determined using a standard am- For this examination, animals were removed plitude-criterion method. In this case, the inten- from their holding pools and anesthetized with sity of the test light at each wavelength was ad- intravenous administration of ketamine (6.0–9.0 justed over successive presentations until it pro- mg/kg) and metatomadine (0.1 mg/kg) inject- duced a response with a constant amplitude of ed transdermally into the dilation of the exter- 3.2 ␮V. nal jugular vein at the base of the head. Once ERG temporal response functions were also anaesthetized, the turtles were positioned on a obtained. To accomplish this, the intensity of an slant board that was placed on an adjustable sur- achromatic (2850 K) flickering light was adjust- gical table, and the head position was stabilized ed to produce a response having a criterion am- through the use of padded restraints. Although plitude of 3.2 ␮V. This procedure was repeated the general anesthetic typically produced some for flickering lights varying in steps of 4 Hz degree of pupillary dilation, in some cases ad- from 4–40 Hz. For all of the spectral sensitivity ditional dilation was achieved by topical appli- and rate measurements, thresholds were deter- cation of atropine and ophthalmic neosyne- mined twice for each test condition and these phrine. The cornea was anesthetized by a topi- values subsequently averaged. All recordings cal application of proparacaine hydrochloride were made under moderate photopic illumina- (0.5%) prior to the installation of a bipolar con- tion produced by a mixture of indirect skylight tact-lens electrode of the Burian-Allen configu- and overhead fluorescent lighting that yielded ration. After the experiments were completed, an illuminance of 165 lux at the subject’s eye. the effects of the anesthesia were partially re- versed using atapamazole (0.1 mg/kg). In all RESULTS cases, full recovery was achieved without inci- dent. As recorded from the corneas of sea turtle The general technique used to record flicker eyes, flickering lights produced small but quite ERGs has been described in detail elsewhere reliable ERG signals. For example, Figure 1A ( Jacobs et al., 1996). In the present experi- shows a representative intensity/response func- 910 COPEIA, 2004, NO. 4 Fig. 1. (A) An example of a flicker ERG intensity/response function obtained from a Loggerhead Sea Turtle. The datapoints are mean amplitudes obtained from four presentations of an achromatic test light flickered at 20 Hz. The intensity of the stimulus is as specified at the cornea.