Yasunori Ishibashi

Three cone opsin genes and cone cell arrangement in retina of juvenile Pacific bluefin tuna Thunnus orientalis

In bluefin tuna aquaculture, collision of juveniles with the tank or net walls is a major cause of high mortality. This problem may be related to color sensibility of the visual mechanisms of this species. As a first step in understanding of color vision of Pacific bluefin tuna Thunnus orientalis, we applied a molecular technique and histology to study cone cell distribution in the retina of juvenile fish. We isolated three cone opsin genes encoding one blue-sensitive (SWS2) and two green-sensitive (RH2) visual pigments. In situ hybridization revealed that SWS2 mRNA is localized in the single-cone photoreceptors. The localization of the two RH2 signals in distinct cone cells was not determined, probably because of the high homology between the two RH2 genes. Single-cone photoreceptors appeared frequently in the ventral-temporal retina in approximately 80-mm fish and in the temporal retina in approximately 230-mm fish. These cone distributions may define a visual field with best color contrast vision in front and above the fish with a short wavelength (blue) reflecting target (sensed by single cones), and may be enhanced against the longer wavelength (green) background when fish see a target below them (sensed by double cones).

Electroretinographic Analysis of Night Vision in Juvenile Pacific Bluefin Tuna (Thunnus orientalis)

We used electroretinogram recordings to investigate visual function in the dark-adapted eyes of the juvenile scombrid fishes Pacific bluefin tuna (Thunnus orientalis) and chub mackerel (Scomber japonicus) and the carangid fish striped jack (Pseudocaranx dentex). Despite the fast swimming speed of the Pacific bluefin tuna, analysis of flicker electroretinograms showed that visual temporal resolution in this species was inferior to that in chub mackerel. Peak wavelengths of spectral sensitivity in Pacific bluefin tuna and striped jack were 479 and 512 nm, respectively. The light sensitivity of Pacific bluefin tuna was comparable to that of chub mackerel but lower than that of striped jack. The Pacific bluefin tuna may not need high-level visual function under dim light conditions in natural habitat because it is a diurnal fish. However, this low temporal resolution and light sensitivity probably explain the mass deaths from contact or collisions with net walls in cultured Pacific bluefin tuna.

Changes in the retinal cone density distribution and the retinal resolution during growth of juvenile Pacific bluefin tuna Thunnus orientalis

Tuna and marlin have well-developed vision. It has been suggested that the vision of the tuna is a major sensory element greatly affecting its behavior. In a behavioral experiment, Nakamura determined the visual acuity of adult skipjack tuna Katsuwonus pelamis. Kawamura et al. calculated the visual acuity of adult bluefin tuna Thunnus thynnus, histologically. Although there are developmental changes in the visual capability and increases in the visual acuity in many fish species with growth, juvenile Pacific bluefin tuna Thunnus orientalis has not been investigated. Furthermore, Fritsches et al. noted that the striped marlin Tetrapturus audax has different visual capabilities along different visual axes. The position of the area of high cell density, or the area centralis in the retina, is related to both habitat and the main visual axis of feeding behavior. Knowledge of developmental changes in retinal topography gives important clues to behavioral changes that occur with growth. To measure the distribution of retinal cone density at each growth stage, it is necessary to investigate specimens from each stage. Recently, full-cycle culture of the Pacific bluefin tuna Thunnus orientalis was achieved in the Fisheries Laboratory, Kinki University, Wakayama, Japan. Using full-cycle cultured specimens of this species, including juveniles, it was possible to investigate the cone density distribution at each stage in this species. In the present study, the authors determined the developmental changes in the cone density distribution, visual axis and minimum separable angle of cones with growth histologically. Specimens of Pacific bluefin tuna from six stages that were full-cycle cultured during 2003–2004 in the Fisheries Laboratory, Kinki University, were used: 30 (total length [TL] = 4.7 cm), 35 (TL = 6.2 cm), 41 (TL = 6.5 cm), 46 (TL = 11.1 cm), 80 (TL = 31.8 cm) days after hatching, and 1 yearold (TL = 102.6 cm). The eyes of the specimens were enucleated and fixed in Bouin’s solution for 24 h. Each retina was then divided into nine regions for 30to 46-day specimens, 25 regions for 80-day specimens and 33 regions for 1-year specimens, according to the eye size. After paraffin embedding, the retina was cut into sections parallel to the retinal surface using a microtome. The sections were stained with hematoxylin–eosin. For quantitative analysis, the sections were examined under light microscopy, photographed, and the number of cones in 0.01 mm of each region in the photomicrograph was counted. Photomicrographs of tangential sections of the retinal cones showed that twin cones form a regular mosaic of parallel rows, although there were occasional irregular single cones in each region from specimens at every stage. An example of a retinal tissue older than 1 year after hatching is shown in Figure 1. The total numbers of twin and single cones in 0.01 mm of each region were counted. Then, the density distributions of both cones determined from the right retinas of the specimens (30, 35, 41, 46, 80 days, and 1 year after hatching) were graded, drawn and displayed using contour lines at 25-cone intervals (Fig. 2). The specimens at even the youngest stage had already metamorphosed and were feeding on fish prey. The characteristics of the cone density distribution changed with each growth stage. In the specimen aged 30 days after hatching, the cone density in each peripheral area reached five times (668/132 = 5.03) that of the bottom region. Therewasnodefinitedirectionof acutevision, such as a visual axis, since no specialized region of maximum cone density existed. It seems reasonable to assume that the distribution at this stage *Corresponding author: Tel: 81-742-43-1511. Fax: 81-742-43-1316. Email: ns_torisawa@nara.kindai.ac.jp Received 10 November 2005. Accepted 27 March 2006. FISHERIES SCIENCE 2007; 73: 1202–1204