Visual ecology in the deep sea: bioluminescence

The deep sea is by far the largest habitat on earth and yet our knowledge about its inhabitants is rudimentary at best, and progressing only slowly. Historic concepts about the deep sea as a lifeless desert have long been proven wrong, just as the common concept about the continual darkness of the abyss. Sunlight plays a minor role between 500 and 1,000m of depth, and is no longer detectable below 1,000m. However, an alternative kind of light is present in the deep sea. Bioluminescence is the major source of light at depths below 200m and found in in most forms of metazoan marine life and across all taxa. Bioluminescence is light that is typically generated by the organism itself, via light emitters known as luciferins (conserved genetically) and enzymes called luciferases (genetically diverse). Bioluminescence appears to be so powerful it is thought to have evolved independently more than 40 times. Residual sunlight and bioluminescence at depth between 500 and 1,000m, also called the mesopelagic habitat, create a visual environment that is markedly different from our world: instead of an evenly illuminated scenery it consists mainly of point light-sources in varying spatial and temporal patterns, vaguely reminiscent of fireworks, which are visible at distances up to about 10m.

Observations in the “wild” from submersibles, and from specimens recovered alive from catches in the laboratory have shown remarkably diverse patterns of bioluminescence. This begs questions about the role of bioluminescence. So far the biological significance of these often highly elaborate light displays is largely a matter of speculation but the probable uses range from camouflage by counterillumination of the ventral side (hatchetfish Argyropelecus, Fig. 1, right), disturbance of predators by release of luminous clouds; communication with, and/or identification of sexual mates; luminous lures to capture prey (anglerfish); illumination of potential prey by “headlight photophores” (some lanternfishes). In general, the wavelengths emitted by the photophores is a bluisch green  ( about 480nm) which closely matches the colour of the downwelling sunlight at mesopelagic depths. In very few cases, however, dragonfish carry light organs under their eyes that emit far red light in addition to the ordinary bluish photophores elsewhere on their bodies. This red light gives them a “private” communication channel and makes them invisible to other animals.


Figure 1: Portrait gallery of mesopelagic fishes with tubular eyes. Note the optical fold in Scopelarchus. The ventral photophores in Argyropelecus serve counterillumination camouflage.

Adaptations of visual systems: Eye designs

Judging by the relative volume of the optic tectum vision plays a major part in the behaviour of mesopelagic animals. Given that residual sunlight plays no major role in the mesopelagic habitat increased visual sensitivity is of utmost importance. Evolution appears to have used two main mechanisms that serve this aim:

(1) Increasing the pupil size to allow for more light to enter the eye.
(2) Optimising photoreceptor sensitivity via several strategies:

(i) The vast majority of deep-sea fish use only rods, the more light sensitive of the two classes of photoreceptor.
(ii) The area of photoreceptive membrane is maximised either by increasing outer segment length to over 100µm or to stack shorter rod inner and outer segments to form multiple banks, giving them an advantage over humans by at least a factor of two.
(iii)  Deep-sea fish rhodopsin’s absorption maximum is tuned to the bluish-green colour wavelength of the downwelling sunlight and bioluminescence, and
(iv) Furthermore, the visual pigment density is unusually high.

Even the early investigators at the end of the 19th century have noted that most deep-sea fish have strikingly large eyes. Demersal (deep bottom living) species are relatively large (up to >1m), and the general body shape is not affected by big eyes. However, fish living in the water column between 500 and 1,000m (called mesopelagic) are considerably smaller (5-35cm) and their large eyes, and consequently large heads may conflict with a more delicate body shape and their ability to swim agilely. Tubular eyes are found in many mesopelagic fish species (Fig. 1) and may be considered as a mechanism to overcome this dilemma. They may be interpreted as cylindrical central segments of the ordinary, hemispherical eyes. This design conserves the large pupillary opening, and the lens projects a focussed image onto the well developed retina at the bottom of the cylindrical eye. The vertical walls of the tubular eye are lined with an “accessory“ retina characterised by thinned layers receiving unfocussed light. The advantage of saving volume comes at a serious drawback, however, being that Tubular eyes have a much-reduced visual field. In most species the optical axes are directed dorsally making them optimally suited to detect silhouettes of animals above blocking the residual sunlight. On the other hand, this eye design would make their bearers vulnerable to attacks from other directions.

Several additional “tricks” have been observed that alleviate this “flaw”: In some species, tubular eyes can be swivelled along a transversal axis extending the visual field rostrally. Other species have developed so called “optical folds”, on the lateral corneo-scleral junction. These are refractive thickenings containing regular layers of precisely ordered and orientated collagen fibrils (reminiscent of the stroma in our own cornea) that result in an extension of the visual field in the latero-ventral direction (see Scopelarchus, Fig.1).

Since we are so used to eyes with lens optics that operate in sun and moonlight to image sceneries as well as details, the diversity of evolutionary adaptations in eye design found in deep-sea fishes with their different optical environment, have fascinated researchers for ages. However, the functions of some structural modifications, though known for decades, have not been understood. On one of my previous deep-sea expeditions with the German research vessel “FS Sonne” we were lucky to make some progress in this respect.

Dolichopteryx longipes: the four-eyed spookfish

 Figure 2: Dolichpteryx longipes, a fish with a conventional tubular eye with refractive optics, and a diverticulum based on reflecting optics. A Dorsal view of a living specimen; the dotted line indicates the plane of sectioning of the following micrographs. B Thin section of the tubular eye and the diverticulum stained with methylene blue. C Diagram of the adjacent section: sclera-blue; choroid-black; retinal pigment epithelium-brown; rod inner and outer segments-green; remaining retinal layers-yellow; mirror-grey. D Visual field angles of the tubular and the diverticular eye. E&F Bright field and dark field micrograph of the diverticulum: Note the bright reflection of the mirror in dark field illumination with polarised light. G Schematic representation of ray tracing modelling demonstrating the function of the focussing mirror of the diverticulum (Fresnel principle) H&I High magnification bright field and dark field light micrographs of the mirror structure showing reflective crystals. J Electron micrograph of stacks of guanine crystals showing alternation of crystals and cytoplasm, each layer about 120nm wide, corresponding to the quarter wavelength of the incoming light and resulting in mirror-like reflection.

In the area of the Tonga trench, in a trawl from between 600 and 800m, we discovered a fish with most peculiar eyes none of our group had ever seen before (Fig. 2A). Preliminary identification put it into the family of barreleyes (opisthoproctids). It had well-developed tubular eyes that were intensely orange when viewed from above (caused by high density unbleached rhodopsin). Next to the tube eye, there was a black circular structure which after turning the fish belly up, showed exactly the same bright orange colour as the tubular eye on the opposite side. We knew at once that this was something very special and took a number of photographic records; these were useful for the later identification as the spookfish Dolichopteryx longipes. However, we had no idea how to interpret what we saw.

We decided to try a dual approach for further investigation. We used the right eye for fine structural analysis, and the left eye, along with the optic nerve and brain for tracer investigations of the visual system. Unfortunately this latter experiment yielded only incomplete results. By contrast, the right eye was serially sectioned, first at about 20µm, and subsequently the more interesting regions, like the retina, were resectioned at 1µm for light microscopy, or at 50nm for electron microscopy.

At low magnification, a section parallel to the optical axis shows a classical tubular eye, with a reduced iris, a spherical lens, a pure rod multibank main retina, and an accessory retina along the medial wall (Fig.2 B,C). Laterally, however, there is a large outpocketing, or diverticulum, enclosed by a sclera continuous with the tube eye (Fig.2 E). Ventrally, this connective tissue layer is markedly thin and transparent, effectively forming a cornea-like structure. Inside the lateral wall of the diverticulum, a typical multibank retina is found, whilst medially, there is a septum-like structure separating the diverticulum from the tubular eye. Intriguingly, this septum shows a prominent thickening that turned out to be hard to focus (Fig.2 E,H), and looked like none of the known layers of the eye or the retina. In order to improve the focus I decided to change the illumination and switched to dark field and polarisation. When I started to rotate the section, major parts of this structure started to light up brightly (Fig.2 F,I) demonstrating that this enigmatic structure was indeed optically active and might represent a mirror built into the diverticulum.

At this point, I consulted with my colleagues and we decided to run an optical modelling programme in order to see whether this structure was capable of reflecting, and possibly focussing light that would enter the diverticulum through the ventral cornea onto the lateral retina. However, for the calculations we first had to further study the fine structure of the presumed mirror. It consisted of stacks of crystals (guanine, in analogy to studies of fish scales and retinal pigment epithelium in other species) between 115 and 125nm thick and separated by layers of cytoplasm of the same width (Fig.2J). Such an arrangement with quarter wavelength stacks of alternating layers (bioluminescence: 480 nm) of different optical densities is well known from other animals to act as a potent reflective device. In addition, we recorded the orientations of these crystals along the length of the mirror and observed that they started almost parallel to the septum ventrally, and then increasingly changed the orientation more dorsally. Modelling these data demonstrated that the incoming light from below was thus effectively reflected and focussed at the outer limiting membrane level, i.e. the rod outer segment acceptance opening of the lateral retina (Fig 2G). By adding such a downward-facing diverticulum to its tubular eye, therefore, Dolichopteryx almost doubles the range of its visual field (Fig 2D) and is capable of surveying the waters above and below.

We have been privileged to be the first to report this mirror based focussing device in the diverticulum of the Dolichopteryx eye, an optical arrangement already known from several arthropods and described by K.. Kirschfeld, one of the founders of the Tübingen Graduate Training Centre and IMPRES for Neuroscience, but never before observed in vertebrates. Interestingly, focussing lenses with a similar optical principle were developed by A. J. Fresnel in the 19th century for early lighthouses.

Barreleyes: a deep sea fish family as an “evolutionary playground“ for different eye designs

As exciting as our discovery of the Dolichopteryx mirror eye was at the time, it turned out to be the key to another surprising story when we looked at the eyes of the other members of the barreleye family. The opisthoproctids comprise seven genera and 19 species; however, the taxonomic relationships are far from settled. All of these have tubular eyes with diverticula of various sizes and degrees of complexities. Two families Opisthoproctus  (Fig.1) and Winteria have tiny outpocketings made up of all three ocular layers, i.e. sclera, choroid and retina, with an unpigmented ventrolateral “window” admitting light to a diverticular retina that lacks ordered, mirror-like crystals.

The intriguing observation here is that in larval specimens of Dolichopteryx, described earlier, the diverticulum looks very similar to this simple, or “primitive” situation in Opisthoprocus. It is tempting to speculate that this might be another case where ontogeny recapitulates phylogeny. A further barreleye species (Bathylychnops) has a diverticulum considerably larger than in the two previous families; however, instead of a crystal mirror it uses a corneal (connective tissue) lens to focus light onto the retina of the diverticulum. Finally, Rhynchohyalus presents a situation at fist sight very similar to the “mirror eye” diverticulum of Dolichopteryx. However, the mirror with its guanine crystals, which in Dolichopteryx is derived from the retinal pigment epithelium, in Rhynchohyalus is made up by iridocytes of the choroid.

These observations lead to at least two further questions:

(i) What is so special about barreleyes that makes them the only family of deep sea fish (known to date) to have evolved such a variety of different eye designs? And
(ii) What are the taxonomic relationships within the opisthoproctid family?

Clearly more morphological data will not help; instead molecular data and markers are needed to solve these questions. And since these require fresh tissue for DNA or RNA analysis, more deep-sea cruises are planned. Apart from the excitement of exploring the fascinating fauna of the deep, the work on a research ship provides an unforgettable social experience: a combination of hard work on board, focussed, and without the distractions of the lab at home, coupled with intense and productive interaction with friends and colleagues.

Prof Dr. Hans-Joachim Wagner works at the University of Tübingen in the Department of Anatomy and enjoys mixing both scientific discovery and thrill-seeking adventure. 

Source of cover image:

Wagner, H.-J., Douglas, R.H., Frank, T.M., Roberts, N.W., Partridge, J.C. (2009). A novel vertebrate eye using both refractive and reflective optics.  Current Biology, 19, 106-114


Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: