We collected Rimicaris exoculata and Mirocaris (Chorocaris) fortunata shrimps at the Rainbow and Lucky Strike sites, respectively, using the submersible Nautile during the AMORES Marvel cruise of RV Atalante in August 1997. We captured the animals by using a suction pump in floodlight illumination and brought them to the surface in a blacked-out Perspex chamber, which provided limited protection against surface light exposure.

The thoracic eyes of some individuals were pink, with a smooth outline and regularly dappled appearance (Fig. 1a), whereas others were a matt chalky white, often with dark areas or streaks in the otherwise featureless reflector (Fig. 1b). We examined the morphology of pairs of specimens of both R. exoculata and M. fortunata with pink and white eyes, each pair taken from the same sample. The pink-eyed specimens show the normal extensive rhabdom (photoreceptor) layer5,6, although there is some evidence (confirmed by electron microscopy) of recent damage to the microvilli (Fig. 1c). The white-eyed specimens of both species show severe breakdown, often with complete loss of the rhabdom layer (Fig. 1d).

Figure 1: Eyes of deep-sea shrimp.
figure 1

a, b, The dorsal surface of live Rimicaris exoculata showing variations in the thoracic eye: a, the pink-eyed type, possibly coloured by rhodopsin; b, the white-eyed type, apparently with only the reflective tapetum. c, d, Sections through resin-embedded specimens of Mirocaris fortunata. c, In pink-eyed specimens there is an extensive rhabdom layer (rl, arrow) extending beneath the carapace from the midline to the lateral margin. d, In white-eyed specimens there is no rhabdom layer (arrow), although the tapetum (t) remains undamaged. Scale bars: a, b, 1 mm; c, d, 100 μm.

The Rainbow site was discovered in 1994 by remote physicochemical sampling without illumination7. The submersibles Alvin and Nautile first visited the active vents in July 1997. We suggest that the retinal damage observed in white-eyed R. exoculata, collected just one month later, was caused by the lights used during these surveys. The shrimps swarm over the vent chimneys and are illuminated by any vehicle working at the active region. The Lucky Strike sites have been visited many times, and we believe that the differences in our Lucky Strike specimens of M. fortunata, most of which had white eyes, have the same cause.

The eye structures of vent shrimps have varying degrees of abnormality, usually ascribed either to poor fixation or to light damage during collection5,6,8. Alvinocaris, for example, apparently has no rhabdoms9, but this may be a consequence of previous encounters with a submersible, rather than being a specific adaptation. The only cases where damage has not been observed are those of juvenile specimens taken by trawling in midwater well above the vents10. These shrimps would not have been subject to previous floodlighting.

The rate of onset of retinal pathology is slow enough for the structure of the retina to be relatively unaffected over a period of hours (as can be seen from our illuminated pink-eyed specimens) but rapid enough for dramatic deterioration to occur in the few weeks between the initial visits to the Rainbow site and our capture of the R. exoculata specimens. There is at present no means of working at the vents without causing this damage, so every vent population visited will already have been exposed to it.

We have established an associative link but not a causal one. Confirmation of these conclusions will require study of the eyes of shrimp captured at the first visit to any new vent site and, ideally, an in situ time series of light-exposed specimens from the same site. Meanwhile, any behavioural observations at previously visited vent sites may relate to shrimp that are already blind.