Prey selection by larval weakfish (Cynoscion regalis): the effects of prey size,speed, and abundance

We examined feeding by larval weakfish, Cynoscion regalis (Bloch and Schneider), in laboratory experiments conducted during the 1991 spawning season. under natural conditions weakfish larval development is ca. 3 wk, and we ran separate experiments with larvae of five different ages (5, 8, 11, 14, and 17 d post-hatching). We used two different size classes of rotifers (Brachionus plicatilis) and brine shrimp nauplii (Artemia sp.) as prey organisms. Contrary to results of previous research, weakfish larvae did not select prey based on size alone. When prey abundance was above 100 itemsl^-1 weakfish, larvae always chose large rotifers (length = 216 m) over small rotifers (length = 160 m). At 11 d post-hatching, larvae switched their diet from large rotifers to small brine shrimp nauplii (length = 449 m); however, when fed small rotifers and small brine shrimp nauplii the change in diet occurred at 14 d post-hatching. This pattern of selectivity was maintained in each larval age class. Early-stage larvae (5 and 8 d post-hatching) did not feed selectively when prey abundance was less than 100 itemsl^-1. Late-stage larvae (17 d post-hatching) fed selectively at abundances ranging from 10 to 10000 items^-1. Lwimming speeds of prey items, which ranged from 1 to 6 mms^-1, had no consistent effect on prey selection. These results suggest that weakfish larvae are able to feed selectively, that selectivity changes as larvae age, and that selectivity is also influenced by prey abundance.     

Performance level and efficiency of two differing predator-avoidance strategies depend on nutritional state of the prey fish

Animal prey has developed a variety of behavioural strategies to avoid predation. Many fish species form shoals in the open water or seek refuge in structurally complex habitats. Since anti-predator strategies bear costs and are energy-demanding, we hypothesised that the nutritional state of prey should modify the performance level and efficiency of such strategies. In aquaria either containing or lacking a structured refuge habitat, well-fed or food-deprived juvenile roach (Rutilus rutilus) were exposed to an open-water predator (pikeperch, Sander lucioperca). Controls were run without predators. In the presence of the predator, roach enhanced the performance of the anti-predator strategy and increased the use of the refuge habitat whereby food-deprived roach were encountered more often in the structure than well-fed roach. Nonetheless more starved than well-fed roach were fed upon by the predator. In the treatments offering only open-water areas, roach always formed dense shoals in the presence of the predator. The shoal density, however, was lower in starved roach. Starving fish in shoals experienced the highest predation mortality across all experimental treatments. The experiment confirmed the plasticity of the anti-predator behaviour in roach and demonstrated that food deprivation diminished the efficiency of shoaling more strongly than the efficiency of hiding. The findings may be relevant to spatial distribution of prey and predator–prey interactions under natural conditions because when prey are confronted with phases of reduced resource availability, flexible anti-predator strategies may lead to dynamic habitat use patterns.     

Effects of nutrition and temperature on metabolic enzyme activities in larval and juvenile red drum,Sciaenops ocellatus,and lane snapper,Lutjanus synagris

The metabolic enzyme activities were determined in larvae of red drum, Sciaenops ocellatus, and lane snapper, Lutjanus synagris, to determine the effect of temperature and nutrition on metabolic enzyme activities and to evaluate if metabolic enzyme activities are useful in assessing the feeding condition of larval fish. During experiments conducted during the spring of 1990, lactate dehydrogenase (LDH) activities in both red drum and lane snapper were approximately an order of magnitude lower than values typical for adult fish; LDH and citrate synthase (CS) activities increased during early developmental stages, but nutritional effects were apparent. Clear differences (up to 4-fold) between well-fed and starving fish were evident in both LDH and CS activity in red drum. Differences between well-fed and poorly fed larvae were evident until 9 d after hatching. Lane snapper larvae reared at a 25°C had significantly lower LDH activities than larvae reared at 28°C.     

Sex,food and conflicts: nutrition dependent nuptial feeding and pre-mating struggles in scorpionflies

Female and male reproductive interests often differ. In species in which matings are accompanied by a transfer of resources valuable for both participants, such as nuptial prey gifts, conflicts may readily occur. Scorpionflies may use alternative mating tactics. One is to offer a prey item (dead arthropod) to females in exchange for mating. This prey gift tactic includes a conflict because a male must decide on whether to offer the gift rather than to fight the female and consume the gift. The outcome may depend on the nutritional status of both males and females. Males may be more willing to give if they themselves are satiated and the condition of the females may influence the payoff from the males’ investment. Similarly, females may be more willing to accept food gifts if they are in poor nutritional condition. In this study of the scorpionfly Panorpa cognata, I experimentally manipulated the feeding history of both males and females. I observed the outcome of the direct interactions that followed when males that were holding prey were approached by females. I found that well-fed males offered the food gift sooner than males in poor nutritional condition that fed extensively on the food item before offering. Female condition had no significant influence on whether prey items were offered by males or accepted by females. I also found that well-fed males rarely searched for prey to pursue the prey gift tactic in courtship. Thus, the prey tactic does not seem to be the males’ first option.     

Effects on larval crabs of exposure to algal toxin via ingestion of heterotrophic prey

We tested whether ingesting toxic algae by heterotrophic prey affected their nutritional value to crab larval predators, using toxic algal strains that are either ingested directly by larval crabs or rejected by them. Ingestion of toxic strains of the dinoflagellates Alexandrium andersoni and A. fundyense by the rotifer Brachionus plicatilis was confirmed. Rotifers having ingested either algal type for five days were fed to freshly hatched larvae of three crab species, with larval survival and stage durations determined. For both algal/rotifer treatments in all three crab species, larvae fed algae directly died during the first zoeal stage, while those fed rotifers that had been fed either algal strain survived to the experiment’s end (zoeal stage 3). Survival was lower, and stage duration longer, for larvae fed rotifers cultured on toxic algae when compared to those fed non-toxic algae. The role of toxic algae in the planktonic food web may be influenced by its direct or indirect ingestion by larval crabs.     

Feeding behaviour and prey size selection of gilthead seabream,Sparus aurata,larvae fed on inert and live food

Prey selection shortly after the onset of feeding by laboratory-reared gilthead seabream, Sparus aurata L., larvae was studied using larvae fed on two types of microcapsule (hard- and soft-walled) having diameters ranging from 25 to 300 m. Preferences between inert food and live prey (rotifers and Artemia sp. nauplii) were also studied. Seabream larvae were able to ingest inert food from first feeding. Larvae of all size classes ingested hard microcapsules with diameters in the range 25 to 250 m. However, larvae with a total length (TL) below 4 mm preferentially selected particles 25 to 50 m in diameter, larvae of TL 4 and 5 mm preferred particles 51 to 100 m in diameter, while larvae above 5 mm TL preferred particles 101 to 150 m in diameter. With soft microcapsules, larvae always preferred particles larger than in the previous case, and above 4.5 mm TL they preferentially selected particles 201 to 250 m in diameter. In addition, the gradual increase of preferred diameters with increasing TL was more pronounced when larvae were increasing TL was more pronounced when larvae were fed on soft particles. Mean values for prey width/mouth width ratios were approximately 0.24 and 0.30 when larvae were fed on hard-walled and soft-walled microcapsules, respectively, irrespective of the absolute value of larval length. When a mixed diet of live and inert food items was offered, live prey were always preferentially selected, even if the prey width/mouth width ratio was apparently not favourable. Only a physical constraint such as excessive prey width could counter this preference for living prey vs inert microcapsules. These results contribute to our knowledge in larval feeding behaviour, especially in the presence of inert food, and represent a fundamental step in developing prepared food for marine fish larvae.     

Diel variability of feeding activity in haddock (Melanogrammus aeglifinus) larvae in the East Shetland area, North Sea

Investigations of factors affecting feeding success in fish larvae require knowledge of the scales of variability of the feeding process itself and the indices used to assess this variability. In this study, we measured short-term (diel) variability in feeding rates of wild haddock (Melanogrammus aeglifinus) larvae four times per day during a 10-d cruise in the northern North Sea. Feeding activity was evaluated using indices of gut fullness, prey digestive state and biochemical measurements (tryptic enzyme activity). The gut fullness and the enzyme activity indices indicated moderate to high rates of food consumption throughout the cruise. Time series analysis of the three indices showed significant diel variability in all indices and enabled identification of significant lags between food uptake and peak digestive enzyme activity. The typical pattern of food consumption and digestion was characterized by maximal ingestion of prey early in the evening (19:00 hrs) and peak digestive enzyme activity at 01:00  hrs. The time scale over which enzyme activities reacted to prey ingestion was ca. 6 h, and is consistent with expectations from controlled laboratory experiments with other larval fish species. Significant diel variability in tryptic enzyme activity suggests that attempts to relate this measure of feeding success to other variables (e.g. food concentrations) should take care to accommodate natural cycles in feeding activity before making statistical comparisons. Received: 29 October 1998 / Accepted: 18 June 1999     

Effects of limiting access to prey on development of first zoeal stage of the brachyuran crabs Cancer magister and Hemigrapsus oregonensis

We tested the influence of limiting access to prey on larval development of the crabs Cancer magister and Hemigrapsus oregonensis by raising their Stage 1 larvae in the laboratory on different prey densities and with various periods of access to prey. Experiments were conducted in 1995 and 1996 at the Shannon Point Marine Center in Anacortes, Washington, USA. Our results show that crab larvae do not require continuous access to prey for optimal development nor do they appear to require light for prey capture. Survival and duration of Stage 1 C. magister fed continuously on only one-fourth the amount of the control density of prey and those fed at the control density for only 6 h per day were the same as for larvae fed continuously at the control density (20 ml⁻¹). Larvae with cyclic access to prey at the control density for 24 h and then starved for 72 h showed significantly lower survival and longer instar duration to Stage 2. Experiments on Stage 1 H. oregonensis which investigated a combination of prey density, period of access to prey and light/dark conditions during feeding revealed that survival decreased with decreasing prey density or with decreasing feeding period, but no differences were observed during periods of limited prey availability as a function of light or dark conditions. Stage duration was not affected by reduced prey density nor by the light/dark condition at the time of feeding, but it was prolonged when the period of access to prey was limited. The period of access to prey did not affect the weight of Day 1 Stage 2 larvae. Larvae fed high densities of prey for 4 h followed by 20 h of reduced-density diet exhibited the same survival and stage duration as controls that were continuously fed high-density prey. Our results define sub-optimal diets that can be used experimentally to determine the nutritional contributions made by naturally-occurring prey organisms during larval development in the two species. In nature, larvae may satisfy nutritional requirements through periodic encounters with dense prey patches during vertical migrations by day or night. Received: 12 August 1997 / Accepted: 5 February 1998     

The effects of zooplankton swimming behavior on prey-capture kinematics of red drum larvae, <Emphasis Type="Italic">Sciaenops ocellatus</Emphasis>

Most marine fishes undergo a pelagic larval phase, the early life history stage that is often associated with a high rate of mortality due to starvation and predation. We present the first study that examines the effects of prey swimming behavior on prey-capture kinematics in marine fish larvae. Using a digital high-speed video camera, we recorded the swimming velocity of zooplankton prey (Artemia franciscana, Brachionus rotundiformis, a ciliate species, and two species of copepods) and the feeding behavior of red drum (Sciaenops ocellatus) larvae. From the video recordings we measured: (1) zooplankton swimming velocity in the absence of a red drum larva; (2) zooplankton swimming velocity in the presence of a red drum larva; and (3) the excursion and timing of key kinematic events during prey capture in red drum larvae. Two-way ANOVA revealed that: (1) swimming velocity varied among zooplankton prey; and (2) all zooplankton prey, except rotifers and ciliates, increased their swimming velocity in the presence of a red drum larva. The kinematics of prey capture differed between two developmental stages in S. ocellatus larvae. Hyoid-stage larvae (3–14 days old) fed on slow swimming B. rotundiformis (rotifers) while hyoid-opercular stage larvae (15 days and older) ate fast moving A. franciscana. Hyoid-opercular stage red drum larvae had a larger gape, hyoid depression and lower jaw angle, and a longer gape cycle duration relative to their hyoid-stage conspecifics. Interestingly, the feeding repertoire within either stage of red drum development was not affected by prey type. Knowledge of the direct relationship between fish larvae and their prey aids in our understanding of optimal foraging strategies and of the sources of mortality in marine fish larvae.     

Relationship between the position preference and nutritional state of individuals in schools of juvenile roach (Rutilus rutilus)

Summary Position preferences of well-fed and food-deprived juvenile roach were investigated in schools of 2 and 4 fish in the laboratory. Food-deprived fish appeared significantly more often in the front position than their well-fed conspecifics. For fish at the same hunger level, individuals at the front of the school had the highest feeding rate. These results represent the first evidence for a relationship between the nutritional state of individual fish and their positions in a school and suggest a functional advantage of the preference. Offprint requests to: J. Krause at the present address     

Sub-lethal exposure to ultraviolet radiation reduces prey consumption by Atlantic cod larvae (Gadus morhua)

Levels of ultraviolet-B radiation (UV-B: 280–315 nm) reaching the earth’s surface have increased over the past few decades due to ozone depletion. It is well documented that exposure to UV-B radiation increases mortality in marine fish larvae. However, few studies have examined sub-lethal effects of UV-B radiation such as, for example, the possibility that it affects prey consumption by fish larvae. Atlantic cod larvae were exposed to a sub-lethal level of ultraviolet-B radiation (UV-B: 280–315 nm) for 15 h. After the exposure, rotifers (4/mL) were fed to cod larvae at three different post-exposure intervals (20 min, 3 and 6 h). Trials were replicated three times for each post-exposure interval. The number of rotifers in the gut and the percent of empty guts (number of fish with empty gut/number of fish examined × 100) were analyzed. Results were compared to those of unexposed fish (the control treatment). UV-B exposed cod larvae had consumed significantly fewer rotifers than control fish at all post-exposure intervals. There was no significant difference in the frequency of empty guts between fish in the UV-B treatment versus fish in the control treatment at any of the post-exposure time points (although the difference at 20 min post-exposure was borderline significant). These observations suggest that cod larvae exposed to sub-lethal levels of UV-B have lower net energy gain which may lead to reduced growth rate and possibly poorer survival.     

Female preferences based on male nutritional chemical traits

Selection favors females that attend to reliable information about male genetic quality and fitness. Male nutritional condition can be a significant sign of mate quality since poor nutrition can be related to reduced sperm quality, low sperm quantity, sexually transmitted diseases, and parasites. We tested whether female Nile tilapia, Oreochromis niloticus, preferred the chemical cues of males that were well fed with high-protein diet over those fed with low-protein diet. Females do not only discriminate between males but also show a preference for well-fed males, discriminating between the odors with respect to nutritional state, suggesting that they were responding to a food-specific chemical cue. It is therefore likely that nutritional condition is related to the production of pheromones in males. Our results suggest that information about male nutritional state can be conveyed in chemical cues and that females attend to these cues during mate choice.     

RNA:DNA ratios and growth of herring (Clupea harengus) larvae reared in mesocosms

Autumn-spawned North Sea herring larvae (Clupea harengus L.) were released in two outdoor mesocosms of 2500 m³ (A) and 4000 m³ (B). The mesocosms were monitored for temperature, salinity, oxygen, chlorophyll a, zooplankton and herring larvae abundance. The density of suitable prey for first feeding larvae (mainly copepod nauplii) was initially low in Mesocosm A (<0.11^-1) compared to in Mesocosm B (>11^-1). Half-way through the experiment the situation was reversed, with higher densities of prey in Mesocosm A (>31^-1) as compared to Mesocosm B (11^-1). The average temperature declined steadily in both mesocosms from 18°C at release to 11–12°C by the end of the experiment 60 d later. The RNA:DNA values of individual herring larvae were related to protein growth rates and temperature adjusted according to Buckley (1984). A corresponding DNA growth index (Gdi) was given as: Gdi=0.68 TEMP+3.05 RNA:DNA-9.92. The RNA:DNA based growth indices were significantly correlated with other somatic growth estimates. The average estimated protein growth rate in the two mesocosms followed the same temporal pattern as the somatic growth rate, but with a lag of 2 d or more. Residual analysis of the regression of ln RNA versus ln DNA also showed the same temporal pattern as the RNA:DNA ratios, but the shift in condition as estimated by this method occurred more in synchrony with the other somatic growth measures. Larvae in Mesocosm A had RNA:DNA values similar to the starvation control kept in the laboratory the first days after release, confirming that larvae in Mesocosm A initially were in poor nutritional condition. On the other hand, the majority of the herring from Mesocosm B were characterised as starving or in poor nutritional condition towards the end of the experiment. The assessment of growth and nutritional condition were in accordance with independent survival estimates which suggested that the majority of the total mortality occurred during the first 15 d in Mesocosm A and there-after in Mesocosm B.     

Use of chemical cues in detection of conspecific predators and prey by newts, <Emphasis Type="Italic">Notophthalmus viridescens</Emphasis>

Summary. Detection of heterospecific predators and prey via chemical cues is well known, but only a few studies have examined the potential for such discrimination in cannibalistic systems. In newts, Notophthalmus viridescens, adults are opportunistic predators of conspecific larvae. I used a laboratory bioassay to determine whether larvae and adults distinguish between chemical stimuli from members of the different age classes. Larvae distinguished between chemical stimuli from larvae and adults by decreasing their activity only following exposure to stimuli from adults. Decreased activity is consistent with an antipredator response in many prey species, including larval newts. In contrast, adults increased their activity and increased time spent in open areas in response to stimuli from larvae, but not to stimuli from adults. Increased activity is consistent with a feeding response; adults also showed increased activity and increased time in open areas in response to chemical stimuli from familiar heterospecific prey (brine shrimp). The proximate cue that allowed the newts to distinguish between the different age classes is not associated with short-term dietary differences because all stimulus animals were fed the same prey. Stimulus strength was controlled by diluting the stimulus solutions according to the volume of the stimulus animal. Therefore, there appear to be intrinsic differences in the chemical signatures of larval and adult newts.     

Carbon content and nutritional condition of sardine larvae (Sardina pilchardus) off the Atlantic coast of Spain

As part of a “European Sardine/Anchovy Recruitment Program” (SARP), sardine larvae (Sardina pilchardus) were sampled off the Atlantic coast of Spain through the spawning season from March to June. The larvae were analysed for carbon and nitrogen content as a measure of nutritional condition and survival potential. There was no significant diel variation in larval carbon content, but there was a small significant diel variation in nitrogen; the absence of a strong diel signal in elemental composition was ascribed to the overnight retention of the gut contents. There was an increase in carbon content with increase in body length which reached an asymptote at ∼40% carbon content at a larval length of 20 mm. It is argued that larvae with a carbon content of <25% of body weight were nutritionally stressed, with the smaller larvae (<10 mm in length) appearing to be more vulnerable to food limitation. Although larvae with the lowest age-specific carbon content (poorest condition) occurred on the cruise with the lowest food availability, there was no consistent relationship between carbon content and food availability. While the successive monthly estimates of carbon content revealed differences in potential recruitment between months, these were not related to the birth-date distribution of the surviving juveniles. Received: 23 November 1998 / Accepted: 3 March 1999     

Free amino acid and protein contents of start-feeding larvae of turbot (Scophthalmus maximus) at three temperatures

The contents of free amino acids (FAA) and total protein, together with growth and gut-content, of turbot (Scophthalmus maximus L.) larvae reared at 14, 18 and 22 °C were studied from first-feeding to approximately 140 effective day-degrees post hatch (D_eff ^∘). Artemia franciscana nauplii and two species of rotifers were used as prey. Protein content accounted for about 42 and 26% of dry body mass in the A. franciscana nauplii and the rotifers, respectively. The FAA pool constituted 5.6 and 4.8% of the total amino acids in the same animals. The dry body mass of turbot larvae was exponentially related to D_eff ^∘. Protein and FAA contents were linearly related to dry body mass, and were independent of rearing temperature between 14 and 18 °C. At the end of the experiment, however, turbot larvae at 22 °C had lower gut content values, retarded growth rates, and decreased FAA contents and concentrations. Thus, at this high temperature, turbot larvae seem unable to catch and ingest sufficient prey, or to sustain an amino acid assimilation rate from the intestine sufficient to meet metabolic demands. Received: 2 January 1997 / Accepted: 25 September 1998     

Effect of delayed first feeding on larval performance of the spider crab Maja brachydactyla assessed by digestive enzyme activities and biometric parameters

The effect of food deprivation on larval performance of the spider crab Maja brachydactyla was studied in terms of survival, moulting capacity, size, weight and enzymatic activities. Five feeding treatments that differed in the initial age of first feeding larvae (fed from hatching, 2, 4 and 6 days post-hatching and unfed) were tested for 20 days. Newly hatched larvae kept without food supply lasted for 10 days and did not moult; with 50% survival observed at 6 days post-hatching. Larvae (zoea I stage) were only able to tolerate 2 days of food deprivation after the onset of exogenous feeding without their performance being compromised. Multivariate analyses suggest that digestive enzyme activities may be good indicators of the nutritional condition of larvae.     

Ontogenetic development of digestive enzyme activities in larval walleye pollock,Theragra chalcogramma

Activities of digestive enzymes trypsin, amylase and lipase in laboratory-reared walleye pollock, Theragra chalcogramma, were measured from hatching to Day 39 (just before notochord flexion) in 1993. All measurements were conducted individually or semi-individually (groups of two larvae of the same standard length). Close relationships between digestive enzyme activities and morphological development of digestive organs were observed. Activities of trypsin and lipase were low during the transition period from endogenous to exogenous energy. Amylase activity was constant with large variance during the same period. Specific enzyme activities of trypsin and amylase indicated high values with large variance during the early period. All three enzyme activities increased with age afterthe transition period, and the specific enzyme activities became constant. The existence of two types of lipase was suggested. One lipase showed a peak of specific activity at Day 4 and might be related to yolk-sac absorption. The activity of the other lipase increased with age after Day 14 and might be related to digestion of prey lipid. Our results suggest that digestive enzymes included in food organisms supplement larval pollock digestive enzymes.     

Food and feeding dynamics of the larval Antarctic fishNototheniops larseni

The feeding dynamics of larvae of the Antarctic fishNototheniops larseni were analyzed from data collected over three years in Bransfield Strait and adjacent waters (Antarctica). Seasonal feeding was examined from 1977/1978 (November–March). The diel feeding cycle was investigated during a 96 h station established in February 1976, while food selection was analyzed using larvae and zooplankton samples collected in February 1982. Hatching occurs in early spring, and larvae fed on eggs of calanoid copepods and on cyclopoid copepods. Copepod eggs were the principal food near the pack ice, and cyclopoids in open waters. Cyclopoids were the staple food in summer. Eggs of the Antarctic krillEuphausia superba were ingested selectively and formed major portions of the larval summer diet in neritic (Joinville Island) and oceanic (Elephant Island) spawning areas ofE. superba. In the fall, copepods predominated in the diets. Most abundant and most frequently ingested prey in summer and fall wereOncaea spp. Feeding commenced at dawn and continued at least until dusk. Krill eggs were taken chiefly during morning hours and egg incidence declined during the day, suggesting that eggs were ingested soon after spawning. Prey size at the onset of feeding was estimated as 0.130 to 0.330 mm. Size-selective feeding was evident in small larvae, while in larger larvae median prey length remained constant. High feeding incidence among yolk-sac larvae in spring, high overall feeding incidence in summer, and size-selective foraging of small larvae suggested favorable feeding conditions in the 1977/1978 season. Yolk-absorption times in Antarctic fish larvae vary on a scale of weeks and may be further retarded due to early feeding. Hence, year-to-year variability of yolk incidence inN. larseni indicated variable biotic environments of early feeding larvae rather than temporal shifts of hatching periods. As hatching periods are constant between years in contrast to the variable retreat of the pack ice and subsequent onset of the production cycle in space and time, maternal yolk reserves are probably utilized to compensate for such variations.     

Depth-independent reproductive characteristics for the Caribbean reef-building coral<Emphasis Type="Italic"> Montastraea faveolata</Emphasis>

Photosynthetic, growth, and reproductive characters for colonies of the massive coral Montastraea faveolata were measured at 3-m-depth and 18-m-depth populations in the Florida Keys in 1993 to investigate reproductive behavior across depth. I tested the hypothesis that reproductive traits would exhibit depth-related decreases similar to those documented for photosynthetically derived energy availability and growth rate. Colonies at the 18 m site had significantly reduced access to photosynthetically derived energy compared to colonies at 3 m depth. Linear skeletal extension rates were also reduced at the deep site (mean±SE; 3 m: 6.13 mm year⁻¹±0.35; 18 m: 3.44 mm year⁻¹±0.36; P<0.001). Contrary to expectations, depth did not affect most reproductive characters. Reproductive allocation, the amount of lipid invested in reproduction, was similar between depths (3 m: 3.74 mg lipid cm⁻²±0.55; 18 m: 3.88 mg lipid cm⁻²±0.62; P<0.8). Reproductive effort, the proportion of total lipid invested in reproduction, was also similar between sites (3 m: 41.4%±4.7; 18 m: 47.3%±7.12; P<0.163). Colonies within a depth did differ significantly in many gametic characteristics including eggs per gonad, total egg area, spermary area, total gonad area, and male allocation (the proportion of total gonad area invested in spermary tissue), but only total egg area differed between depths, and this was greater at 18 m (3 m: 55.63 μm²×10⁴±4.76; 18 m: 60.51 μm²×10⁴±3.41; P<0.035). Colonies at the two sites had similar polyp fecundity (egg production per polyp); however, a difference in polyp density between sites (3 m: 5.56 polyps cm⁻²±0.21; 18 m: 4.87 polyps cm⁻²±0.15; P<0.028) resulted in a significant reduction in area fecundity (egg production per square centimeter) at the 18 m site (3 m: 554 eggs cm⁻²±27.9; 18 m: 422 eggs cm⁻²±31.0; P<0.045). Thus, there was a depth-related fecundity cost arising from a change in colony architecture (i.e. polyp spacing) rather than a direct effect of depth on the number of eggs produced per polyp. Egg production per gonad was positively correlated between 1993 and 1994 (r=0.540, P<0.05), indicating that other reproductive characteristics may also be stable between years.