Research ships are much more numerous on the high seas in the 1990s than they were in the 1950s. At that time the United Kingdom was fortunate in that it had a purpose-built research vessel, RRS 'Discovery II' (see inside back cover), the successor to Scott's 'Discovery' (which was itself granted the title Royal Research Ship in 1923, see inside front cover). 'Discovery II's prime objectives had originally been oceanographic research in the Antarctic relevant to the whaling industry, to which end she had completed six two-year Antarctic commissions. In the 1950s she was being operated by the National Institute of Oceanography and was working extensively in the North Atlantic, covering the gamut of marine physics, chemistry, biology and geology, including participation in the research programme of the International Geophysical Year. However, by 1959 she was 30 years old, and both scientific space and stability had become limiting. Increasingly stringent Ministry of Trade requirements were also making successive refits less and less cost-effective. In April of that year the first brief outline for a replacement vessel was prepared, and six months later discussions between naval architects and the National Oceanographic Council (represented by Dr H.F.P. Herdman) produced an outline draft specification for the new ship. Further discussions with the Admiralty finally resulted, in June 1961, in an order being placed for a new vessel, at an estimated cost of £800K (Deacon, 1967).

The new ship, the RRS 'Discovery', was launched in July 1962 and commissioned on 17 December, 10 m shorter than her originally conceived length but at a contained final cost of £802K.

Three latitudinally separated assemblages (49°–41°N, 38°–30°N, 22°–20°N in 11°–25°W) of abyssal demersal fishes in the eastern North Atlantic have been identified from combined data from 13 new and 12 previously reported trawl samples. The northerly assemblage remained unaltered by the new data, but the intermediate one could be sub-divided into shallow (4370–4475 m) and deep (4835–5440 m) elements. This sub-division revealed that bathymetry was acting with latitude to influence previous results. The new, southerly, assemblage was found below surface waters enriched by upwelling along the continental margin. The composition of dominant species identified each assemblage. The biological characteristics of these species were largely consistent with the former hypothesis linking them to the influence of the overlying production regime. The contrast between this preliminary evidence for structured demersal fish assemblages in the abyss with its lack on the slope and rise was discussed in the terms of the respective patterns of trophic input.

‘Four-winged’ flying fish (in which both pectoral and pelvic fins are hypertrophied) reach greater maximum sizes than ‘two-winged’ forms in which only the pectoral fins are enlarged. Exocoetus obtusirostris shows negatively allometric growth in relation to standard length in terms of body mass (b=2·981), and lateral fin area (b=1·834). In consequence, wing-loading rises in positive allometric fashion with standard length (b=l·236). Pectoral fin length cannot be greater than 78–79% of standard length or swimming will be impaired, so the requirement for increased flying speed resulting from increased wing-loading during growth means that lift:drag ratios have to be improved by relatively narrowed wings and tapered wing tips; features which in turn increase wing-loading. Evidence is presented to show that hypertrophied pelvic fins in four-wingers are required to solve problems of stability in pitch, rather than to decrease wing-loading. The ‘non-flying’ flying fish, Oxyporhamphus micropterus, has very high wing-loadings, but the main reason that it cannot fly is that the centre of gravity of the fish is so far behind the pectoral fins that stalling on take-off would be inevitable. Flying fish possess reasonable quantities of red axial musculature, but no more than are used for cruising in fast-moving pelagic fish such as mackerel; it seems probable that acceleration to take-off speed in flying fish requires use of anaerobic white muscles.

Video recordings were made of the behaviour of hyperiid amphipods and other mid-water crustaceans in a tank lit from above with a dim VDU screen, and from behind with infra-red light. Computer-generated stimuli could be moved across the VDU screen. Amphipods swam at speeds up to 19 body lengths per second. Some (Brachyscelus, Phrosina) consistently swam inverted, rolling from side to side. Others (Platyscelus, Parapronoe) spent periods of time rotating in the water with the dorsal acute zone of the eye pointing upwards. Both behaviours are interpreted as forms of visual scanning. Different species were active at different light levels, Phronima preferring the dimmest illumination. Three kinds of visual behaviour were seen consistently. The animals dropped from the surface as a dark square passed above them; they tracked other individuals (Hyperiids) or, in the case of Euphausiids, the stimuli on thescreen above; and they occasionally swam towards each other and collided, apparently deliberately.

The anatomy of the eyes of several species of mesopelagic decapods (family Oplophoridae), obtained from the eastern north Atlantic, is described and related to the unique light environment of the deep seas. The oplophorid eyes are of the reflecting superposition type, but they show a number of regional morphological variations. The main rhabdom, formed by retinula cells Rl to R7, comes in a variety of shapes, from fusiform rhabdoms in the dorsal region of the eyes of Oplophorus spinosus to multi-lobed interdigitating rhabdoms in deepwater species. The distal rhabdom, contributed to each ommatidium by retinula cell R8, gradually increases in size towards the ventral part of the eye in Systellaspis debilis and O. spinosus. Histological examination of the tapetum shows that it is incomplete dorsally in some species from the upper mesopelagic zone (S. debilis, O. spinosus), and that the amount of reflecting pigment in the tapetal cells increases in the ventral part of the eye. The tapetum is complete in some deep-water species (Systellaspis cristata, Acanthephyra kingsleyi, A. pelagica). These adaptations of the rhabdoms and tapeta are thought to be concerned with increased sensitivity to the dim up-welling irradiance and to bioluminescence. A dorsal accessory compound eye consisting of a small group of apparently functional apposition-type ommatidia is described.

Eyeshine brightness was measured in a number of species of oplophorid shrimps and one sergestid. Eyeshine varies in a systematic way across the eye. Forward and downward looking parts of the eye often have the brightest eyeshine. In many cases eyeshine is graded in the horizontal and vertical axes. Eyeshine brightness can be explained in terms of underlying tapetal morphology. These features of the decapod eye can be rationalised in terms of the normal irradiance distribution in the sea and the need for certain parts of the eye to have enhanced sensitivity. Contrary to the expected results, when six species of oplophorid were compared there was a clear trend of decreasing brightness with depth at whichthey occurred. In a number of species the tapetum is incomplete with distinct holes. Such holes occur in the dorsal region of two species of oplophorid. In one deep-water species the central tapetum is lacking. This feature, and the observed decrease in eyeshine with depth, may be adaptations to reduce the signal that eyeshine provides to potential predators.

The elongated eye of Vitreledonella richardi is described from two specimens caught at 200 m in the North Atlantic. It is shown that the cylindrical shape, with the lens placed centrally, results in a reduced field of view in the horizontal plane. It is argued that the unusual shape is an adaptation which minimises the silhouette of the eye as seen from below, and is part of the animal's camouflage strategy.

Introduction

The majority of fish and cephalopods have two hemispherical eyes, each of which has a field covering roughly the 180° on one side of the animal (Walls, 1942; Packard, 1972). The commonest exceptions to this rule are animals with so-called ‘telescopic’ eyes, found in many species of deepsea fish (e.g. Benthalbella, Argyropelecus, Opisthoproctus, Scopelarchus, see Marshall, 1954; Locket, 1977) and some cephalopods (Amphitretus, see Chun, 1915; Histioteuthis sp., see Young, 1975a). These eyes are tubular, with the lens at the open end of the tube, and a retina with a reduced field of view at the other. The eyes usually point upwards, and it is generally agreed that their shape is determined by the need to concentrate sensitivity and resolution into the rather narrow field above the animal where the residual daylight is still adequate for visual detection. It is 200 times darker in the downward direction (Denton, 1990), and so there must be some depth at which upward vision works, but downward vision does not. These are not the only kind of tubular eyes, however. In some cephalopods, but not to my knowledge in any fish, there are eyes of tubular appearance but with the lens not at the end, but in the centre of the tube (Figure 1).

This study describes the spectral characteristics between 250 and 700 nm of lenses from 71 species of deep-sea teleost. The majority of these did not contain detectable amounts of short-wave absorbing pigment and transmitted all longer wavelengths to the same extent before cutting off smoothly and rapidly in the near-UV (50% transmission 310–348 nm). The lenses of 15 species, however, contained significant amounts of short-wave absorbing pigment. The majority of these lenticular pigments had smooth ‘bell-shaped’ absorbance spectra with a single peak in the near-UV. Such ‘simple’ pigments were extracted from four species (λmax: 360·8 nm, 36·50 nm, 395·7 nm, 324·8 nm) and detected in a further eight species, although insufficient material was available to allow their extraction from the latter group.

The absorbance profiles of lenses from another three species were more complex, suggesting the presence of one and two carotenoid-like pigments respectively in the lenses of Argyropelecus sladeni and A. affinis, and two as yet unidentified pigments in the lenses of Malacosteus niger. This brings the total of short-wave absorbing pigments so far described in the lenses of deep-sea fish to at least eleven. In Argyropelecus spp. the pigmentation was most pronounced in the lens cortex, while in Malacosteus niger it was restricted to the lens core. In all other species the pigment appeared freely diffusible throughout the lens. Several species showed an increase in the degree of lenticular pigmentation in older animals and two (Argyropelecus affinis and Malacosteus niger) demonstrated age-related changes in the relative concentrations of two different pigments within the lens.

The visual pigments in the retinal rods of 17 species of deep-sea fish were examined by microspectrophotometry or visual pigment extract spectrophotometry. In 15 species single visual pigments were found with peak sensitivities between 470 and 490 nm, typical of deep-sea fishes. However, in one species, Stylephorons cordatus, two visual pigments were found with λ values at 470 and 481 nm. In another species, Scopelarchus analis, three visual pigments were found with mean λ values of 444, 479 and 505 nm. The short-wave pigment of this species was found both in main and accessory retinae. It was present both in single rods and in outer segments which had the most long-wave sensitive pigment in their distal parts. It is argued that these two-pigment rods are in the process of changing their visual pigment from a ‘juvenile’ VP505 pigment to an ‘adult’ VP444 pigment. The VP479 was found only as a single pigment in rods in the accessory retina.

Video recordings of bioluminescence have depended on the use of image-intensified systems, which generally suffer from poor resolution and image persistence. These characteristics, coupled with the restriction that any illumination must be dimmer than the bioluminescence, have complicated efforts to record behaviours associated with bioluminescent displays. The Mixed Light Imaging System (Mlis), described here, uses (1) an intensified video camera, (2) an infra-red (IR) sensitive video camera (3) a dichroic beam splitter, (4) strobed IR illumination and (5) a video mixer to record stop-action images of organisms superimposed on intensified images of their bioluminescence. With the MLIS, behaviours associated with bioluminescent emissions, including rapid escape responses, were recorded from the mesopelagic copepods Euaugaptilus magnus and Gaussia princeps.

The pyloric caeca and posterior intestine of Howella brodiei are bioluminescent, and are associated with internal reflective tissues. The bioluminescence is not bacterial in origin. Unlike that of several shallow water apogonids, it does not depend upon Vargula luciferin. Florenciella lugubris has bioluminescent oesophageal diverticula and is able to eject a pulse of luminescent material from beneath the operculum.

The luminescence of the two species is compared with that of other perciform fishes.

Introduction

The perciform family Apogonidae (cardinal fishes) contains several bioluminescent coastal species in the Indopacific genera Siphamia, Rhabdamia, Archamia, and Apogon. Their oceanic relatives are less well known and their systematic position is still a matter of some conjecture. Howella has been assigned to the Apogonidae but has also been included in the Moronidae (Tortonese, 1986), and, more recently, in the Percichthyidae (Post & Quéro, 1991). Some authors have followed Mead & De Falla (1965) in assigning the oceanic apogonid-like fishes to the family Cheilodipteridae (e.g. Fedoryako, 1976), in recognition of the fact that they ‘are so different from their coastal relatives that comparison is pointless’ (Mead & De Falla, 1965).

Bioluminescence has not been described in detail in any of the oceanic species. Mayer (1974) reported that one of the pyloric caeca in Epigonus macrops was separated from the others and associated with reflective structures. This strongly suggested that it was bioluminescent, based on comparisons with known luminescent shallow water perciforms, such as species of Pempheris and Parapriacanthus. Mead & De Falla (1965) described a ventral reflective strip in Rosenblattia robusta which, by analogy with similar appearances in luminous neritic species, they considered a possible indication of bioluminescence.

Sequences of the 16S ribosomal RNA gene of luminous bacterial symbionts from the escas of the deep sea anglerfishes Melanocetus johnsoni and Cryptopsaras couesi were determined by direct sequencing of polymerase chain reaction products. A sequence was also obtained from a strain of Photobacterium phosphoreum, the culturable light organ symbiont of Opisthoproctus grimaldii. Comparison of these and other published sequences showed that the anglerfish symbionts group with the marine luminous bacteria but are not closely related to P. phosphoreum. The two ceratioid symbionts differ from each other at least at the species level.

This paper describes the locomotor movements of the tadpole larvae of the ascidians Ciona and Dcndrodoa and the associated electrical activity of the caudal muscle cells. Although very different in size and trunk shape, both larvae show essentially the same two movement patterns; symmetrical swimming and asymmetrical tail flicks. Swimming movements at tailbeat frequencies up to 40 Hz and forward speeds up to 10 L s (in the Reynold's number range 5·25) involve large lateral movements of the tail, and large yaw of the trunk. Tail oscillations are produced by muscle cells arranged in three segmented rows along the tail, coupled by gap junctions. Swimming is driven by axons innervating anterior ventral muscle cells. The second type of movement, single or multiple tail flicks, is driven by axons innervating dorsal muscle cells, all of which are innervated along the length of the tail. The middle row of muscle cells is not innervated. This small scale oscillatory swimming system is compared with those of chordates and larvacean tunicates, and it is concluded that both of these are very different from that of the ascidian tadpole.

In this paper we present data on the organization of the territories in males of Lipophrys pholis during the breeding season. Data were collected during high tides by skin- and scubadiving and during low tides by direct inspection of nests. Our study area was located at Arrébida, Portugal. The main results are: (i) The territories of the breeding males of this species are temporary, being established each breeding season, (ii) The guarding males stay in the nest holes with the egg masses while the tide is low and are subjected to several hours of emersion in each tidal cycle, (iii) Even when the nests are submerged the fishes stay inside the nest for an average of 92% of the time. All the activities performed outside the nest correspond to an average of 27 minutes per day. (iv) There was a low frequency of territorial intrusions. Conspecific intruders released a significantly higher frequency of agonistic responses than did Coryphoblennius galerita. (v) Removal experiments showed that vacated territories are not occupied by other males during the same breeding season, (vi) Without the presence of the guarding male the eggs are slowly destroyed by predation and infection, but some eggs can still survive and hatch up to five days after the removal of the parental male. The results are discussed in terms of the probable costs and benefits of breeding intertidally.

The distribution of bass, Dicentrarchus labrax (L.), eggs taken in ichthyoplankton surveys of the Bristol channel and eastern Celtic Sea from April 1989 to May 1990 indicates that bass spawned predominantly offshore during March and April. Seventy-two bass larvae were captured by ring net and 16 by high-speed tow net during 19–27 May 1989, and 11 were captured by ring net during April and May 1990. Larvae were most abundant inshore where the water column was unstratified and less than 50 m deep.

Back-calculated egg fertilization dates were determined for 58 larvae captured in May 1989 by ring net, using estimates of temperature-dependent egg and larval development rates and counts of daily growth increments on sagittal otoliths. These dates ranged from 5 April to 10 May 1989, later than those for most young-of-the-year bass which first appeared in Welsh estuaries during June 1989. This implies that bass larvae would have been more abundant before the May sampling period, even though these catches are the largest reported from UK waters.

Bass larvae of 5–11 mm live notochord length were captured close to estuarine nursery areas, and their mean growth rate was approximately 0·2 mm per day. However, the smallest bass first arriving in the nursery areas during June, July and August were always larger than these larvae, and predominantly 15–20 mm total length. It is suggested that bass larvae hatching offshore either perish or are transported to unstratified coastal waters where they feed and remain for at least 30 days prior to their recruitment to the nurseries.

The variation in species composition of continental slope fishes as determined by the catches of different trawls towed either on single or paired warps was analysed by Detrended Correspondence Analysis (DCA). The catches of two trawls, a semi-balloon trawl (OTSB) and a Granton trawl (GT) were very similar when towed on paired warps. Significant differences were found between the catches of the OTSB trawl towed on single and paired warps. The DCA effectively provided information on the important gradients, e.g. depth, trawl type, and indicated which species were most abundant in the different depth zones and trawl types. Detailed accounts of the abundance and biomass of different species by trawl type and depth zone are given and the observed pattern of distribution discussed.

Shell morphological features of the larvae and early postlarvae of species from two venerid (Mollusca: Lamellibranchia) subfamilies, Chioninae (Mercenaria mercenaria, M. campechiensis, M. campechiensis texana and Chionecancellata) and Pitarinae (Pitar morrhuanus), were examined using scanning electron microscopy. A single elongated fold on the anterior half of the left valve became more robust and a small fold in the centre of the right valve increased in length along the hinge line throughout the larval period. A MANOVA of shell length, shell height and provinculum length detected statistically significant (P) differences between the larval shells of Chioninae and Pitarinae, and within the Mercenaria.

Recoveries of tetracycline-labelled specimens of the sea urchin Echinus esculentus (Echinodermata: Echinoidea) from a wild population marked two years previously indicate very low skeletal growth rates in large adults. The post-tag growth in the test of a smaller specimen showed two clear growth zones in the middle layer of the plates, this conforming to the expectation of a single growth band each year. Merging of the spinochrome pigment bands present in the outer layer near the plate edge in older urchins will probably result in underestimation of age based on counts of these bands.

The large literature on growth banding in the European sea urchin Echinus esculentus L. and other echinoids is reviewed by Pearse & Pearse (1975), Smith (1980) and Gage (1991). Moore (1935) utilised spinochrome pigment banding in the genital (apical) plates of E. esculentus from the Isle of Man (Irish Sea) and Firth of Clyde (western Scotland) in one of the first studies utilising growth bands to interpret the age structure and growth rates of sea urchins. A single band was assumed to be formed each year. Counts of spinochrome bands have been used to obtain nearly all subsequently published age data for this species (Sime, 1982; Nichols et al., 1985; Sime & Cranmer, 1985; Comely & Ansell, 1988).

The present study was aimed at helping to resolve differing interpretations of age and growth rates in Echinus esculentus provided by these studies. This was undertaken by time marking the skeletal plates of a large sample of a wild population accessible by scuba diving on a submerged rock reef at 10–15 m depth off the islet of Eilean Mor near the Dunstaffnage Laboratory.

Spherical, opaque bodies (5–10 µm diameter) were frequently observed adhering to cells of the marine ciliate Mesodinium rubrum, particularly towards the end of red-water blooms in Southampton Water. Transmission electron microscope sections revealed these structures to be of biogenic origin, and packed with an amorphous, lipid-like material. No diagnostic features were observed to confirm their identity.

The phototrophic ciliate Mesodinium rubrum (Lohmann) forms recurrent red-tides each summer and autumn in the Southampton Water estuary, UK (Williams, 1980; Soulsby et al., 1984). During a study of the impact of these blooms on the estuary, live cells were routinely observed microscopically. Frequently, but predominantly later in the bloom, numerous individuals posessed spherical, colourless, but opaque bodies, apparently attached or adherent to the outer cell membrane. Cells observed prior to the bloom, or at other times of the year, were never observed o possess such structures.

Cells bearing these bodies were photographed live, or after preservation in Lugol's iodine, using the phase contrast optics of an Olympus BH2 photomicroscope.

The ciliate list for Plymouth waters has been extended by 14 species using modern taxonomic techniques. Ciliates were abundant in the plankton where they formed a significant food resource. Their community biomass and production was estimated to average 12 µ C 1 and 9 µ C 1 respectively during the summer. The ciliate community was dominated by a diverse assemblage of aloricate choreotrichs, suggesting a complex trophic role for this protozoan group.

Ciliate protozoans are ubiquitous and often abundant in marine waters where they are frequently considered to play an important ecological role in trophic flux and nutrient cycling within the plankton (Fenchel, 1987). In spite of this, however, their ecological role in British coastal waters is poorly understood. In Plymouth waters, for instance, there has been only one previous study of marine pelagic Protozoa (Lackey & Lackey, 1963), despite the presence of a marine laboratory in the region for over 100 years. As the study by Lackey & Lackey (1963) focused solely upon the taxonomy of local protists, the ecological role of protozoans in Plymouth waters is unknown. To redress this anomaly the present pilot study was undertaken in Plymouth waters with the following objectives: to identify the dominant ciliates from this region using techniques unavailable to Lackey & Lackey (1963), to quantify ciliate abundance and cell sizes, and to estimate their biomass and production.

Triplicate water samples were collected, using a 3-litre water bottle, from surface waters at each of four stations along a 20-km transect between Plymouth Sound (50°21'N 04°09'W) and the Eddystone Rock (50°ll'N 04°16'W) during June, July and August 1988.