Clusters of barnacles were collected from the SW coast of Portugal (Cabo Sardão, Portugal, 37°36′24.70″N, 8°49′2.00″W) during the reproductive season and transported within 36 h to the School of Marine Science of Technology (Newcastle University, UK), where the barnacles were dissected (n≥60 individuals) to extract egg lamellae. Lamellae were separated according to developmental stage and mature lamellae were cut to assist naupliar release. Nauplii were selected by positive phototaxic behaviour and transferred into culture. Cultures were semistatic, with minimal bottom aeration, in plastic carboys containing 10 L of 0.2 µm natural filtered seawater (FSW) at 20±1°C, 16:8 L/D (dim light, 200-1500 lux), 33±1 psu and antibiotics (0.0232 g l^–1 penicillin and 0.0369 g l^–1 streptomycin), with 1 larva ml^–1. Cultures were fed after water changes every two days with a 1:1 mixed diet of Tetraselmis chuii and Isochrysis galbana (10⁴ cells ml^–1). Larval cultures were monitored daily in order to determine when the majority had reached the cyprid stage, whereupon cultures were filtered through a 60-µm mesh.

A series of six experiments (Exp. 1 to 6) were conducted to test the factors that mediate cyprid settlement on adults and to investigate the effects of abiotic and biotic factors on attachment and metamorphosis. These included testing various environmental conditions, specifically hydrodynamics (Exp. 1; August 2012), temperature (Exp. 2; June 2012), salinity (Exp. 3; June 2013) and light (Exp. 4; August 2013), as well as cyprid age (Exp. 5; August 2011) and batch (Exp. 6; June 2011, 2012, 2013). Two additional experiments (Exp. 7 and 8; respectively done in August 2011 and 2012) were conducted to follow cyprid settlement (attachment and metamorphosis) and survival over time. Identical assays were conducted for all experiments with the setup adjusted for the factor under investigation, as detailed below. Cyprid settlement assays were no-choice assays (except Exp. 4) in 500-ml conical flasks, 0.2 µm natural FSW, kept at 20±1°C, 16:8L/D (dim light, 500-1000 lux), 33±1 psu, antibiotics (as previously) and low bottom aeration. To exclude the possibility of cannibalism, adults were placed in 50-mL Falcon tubes^® (whose lid had been cut to allow the stalk through and sealed with Parafilm), with the stalk towards the inside of the tube and the capitulum on the outside. The tubes were filled with FSW and 50 cyprids were added to each replicate tube (1 larva ml^–1). The tubes were then sealed and placed inside conical flasks, to allow the adult to be submerged and extend its cirri. The larvae could swim inside the tube, with the adult stalk and the inner surface of the polypropylene tube as alternative settlement substrata (Fig. 1). Each treatment comprised 9 adults, one per tube assembly, with 50 cyprids from the same batch (except for Exp. 6, in which cyprids were from different batches). Prior to the experiments, all adults were checked for recruits and the stalks were gently cleaned, to ensure comparable experimental conditions and no previous settlement. Except for data collection, the setup was left undisturbed, with no further water changes or feeds.

Results were analysed using Statistica^®, with a significance level of 0.05. Data analysis was carried out on the basis of homogeneity of variances (Levene’s test) and normality (Kolmogorov-Smirnov test), while significant differences were detected by using one-way ANOVA, followed by a post-hoc Tukey HSD test when relevant. Data as percentages were arcsine transformed pre-analysis. When only two treatments were being compared (e.g. light-shade choice assay) a Student t-test was used for dependent or independent samples, as appropriate. Regressions were used to analyse effects over time on permanent attachment, metamorphosis and survival of the larvae. The curves presented were selected by significance (F and p-values) and best fit (R2) within the measured range and are not a basis for extrapolation outside of this context. All results are presented as the mean ±SD, unless stated otherwise.

Permanent attachment and metamorphosis (settlement) of larvae on adults, according to the different environmental conditions of hydrodynamics, temperature, salinity, light and the cyprid-related factors of age and batch, are shown in Figures 2 and 3 (A to F). Permanent attachment varied significantly with hydrodynamic conditions (ANOVA; F=14.26, p<0.01; Fig. 2A), being lower in the static condition (12.67 ± 4.87%) than in the other treatments (29.93±9.13%; Tukey’s test; p≤0.01), which were not significantly different from each other (Tukey’s test; p≥0.17). Larvae maintained in static conditions were less active than those kept in hydrodynamic conditions. Larvae in the interface treatment tended to swim close to the water surface and settled on the adult stalk just beneath the meniscus. The percentage of larvae metamorphosing within one week of adding cyprids was not affected by hydrodynamic conditions (ANOVA, F=0.29, p=0.83; Fig. 3A), averaging 73.81±10.02%.

Though cyprids kept at higher temperatures were considerably more active than those kept at lower temperatures, no significant differences were observed for permanent attachment (ANOVA, F=0.30, p=0.75; Fig. 2B), which averaged 32.30±12.36% across treatments. However, total metamorphosis over the first week varied significantly according to treatment (ANOVA, F=9.01, p<0.01; Fig. 3B), increasing in line with temperature from 59.78±11.43% at 14°C to 79.22±7.33% at 20°C (significantly different from each other; Tukey Test, p<0.01).

Salinity affected both permanent attachment (ANOVA, F= 5.68, p<0.01) and metamorphosis (ANOVA, F=13.36, p<0.01) (Figs 2C and 3C). Permanent attachment was significantly lower at 20 psu (23.78±5.93%; Tukey’s test, p≤0.04) than at 30 and 40 psu (on average 31.56±5.52%). The same pattern was observed for metamorphosis: 52.11±8.51% at 20 psu and an average of 69.11±8.37% at higher salinities (30 and 40 psu; Tukey’s test, p<0.01).

Cyprid permanent attachment in the light was approximately double that in the dark (Table 1: light vs. dark–no choice; t-test-independent samples, df=16, p<0.01). This preference for attachment in light conditions was confirmed in the choice assay (Table 1: light or shade choice; t-test-dependent samples, df=8, p<0.01), where of the 33.11±6.51% of cyprids that attached, 26.11±6.13% did so on the illuminated side and 7.00±1.93% on the shaded side. No differences were observed in metamorphosis within the first week between different light conditions (Table 1; no-choice, t-test-independent samples, df=16, p=0.67; choice, t-test-dependent samples, df=8, p=0.48), which averaged 72.82±6.92%.

The age of cyprids (Figs 2E and 3E) significantly affected permanent attachment (ANOVA, F=8.60, p<0.01) and metamorphosis (ANOVA, F=5.47, p=0.01). Larvae in all groups appeared healthy, with visible numbers of lipid droplets. Permanent attachment was higher for aged larvae (32.89±5.09% for 3-day-old and 36.89±3.59% for 6-day-old larvae; not significantly different; Tukey’s test; p=0.30) than un-aged larvae (26.44±6.84%; Tukey’s test, p≤0.04). Successful metamorphosis within the first week decreased with ageing and was significantly lower for cyprids aged for 6 days than for those aged for 3 days (60.89±7.83%; Tukey’s test, p=0.01).

Permanent attachment (ANOVA, F= 4.12, p=0.03) and metamorphosis rate (ANOVA, F=7.36, p<0.01) varied with larval batch (Figs 2F and 3F). These differences were significant between batches A or B and C for permanent attachment (Tukey’s test, p=0.04) as well as between B and A or C for metamorphosis (Tukey’s test, p≤0.01); all other comparisons were not significantly different (Tukey’s test; p≥0.08). Permanent attachment varied between 27.00±6.48 and 34.11±5.06%, while metamorphosis ranged between 55.22±7.83 and 69.33±9.28%. It is important to note that the measure of metamorphosed larvae is always expressed as a percentage of the total larvae attached at 48 h and not as the percentage of larvae added initially. Batch differences accounted for variations of the order of 7% and 15% for permanent attachment and metamorphosis, respectively, for groups of larvae reared under similar conditions but collected in consecutive years.

Rate of settlement and cyprid survival over time

Significant differences were recorded for permanent attachment with time (Fig. 4A; polynomial regression: y=–0.0043x²+0.74x–0.33, xϵ[0,96]; R²=0.99, F=133.67, p<0.01). Attachment increased over the first 48 h, after which it stabilized at around 30%. Similarly, the rate of metamorphosis also changed with time (Fig. 4B; polynomial regression: y=–0.54x²+13.34x–1.94, xϵ[0,12]; R²=0.98, F=48.76, p=0.02), increasing to 80.67±4.97% by day 12. About 10% of cyprids failed to metamorphose even after 18 days.

When larval survival in culture was evaluated in the absence of settlement substrata, cyprids lived for extended periods, with mortality after 20 days at around 50%. Survival decreased significantly with time (Fig. 5; polynomial regression: y=–0.09x²+0.69x+100.34, xϵ[0,20]; R²=0.99, F=1479.23, p<0.01). Survival for the first 6 days post-metamorphosis to the cyprid was over 95%, thereafter declining gradually until day 16 (to ≈70%), at which point a steeper decline commenced (Fig. 5). Cyprid activity also declined with time, as did visible lipid reserves (oil droplets). Living cyprids were observed even after 30 days in culture, although the amount of available lipid was virtually non-existent and cyprid activity was minimal.

Environmental conditions (hydrodynamics, light, salinity and temperature) and their effects on settlement were investigated. The effects of water movement on barnacle settlement have been widely studied (e.g. Crisp 1955, Mullineaux and Butman 1991). In the present study, permanent attachment was lower in static conditions, suggesting that dynamic conditions favour settlement. This is in agreement with the conditions experienced by larvae in the natural environment, where flow is highly turbulent. Though the method of water agitation used in this study either facilitated (e.g. aeration) or hindered (e.g. orbital shaker) the circulation of cyprids, no significant differences in attachment or metamorphosis were found according to the type of hydrodynamism, supporting the notion that transport of cyprids to a suitable surface is more important than the flow characteristics. The lack of a difference in the rate of metamorphosis according to hydrodynamic conditions further implied that there was no significant difference in energy expenditure in each treatment. Interestingly, in the air-water interface condition, cyprids were observed at the water meniscus and out of water stuck to the surface. When observed under the stereomicroscope, P. pollicipes cyprids that had settled out of the water appeared to be covered by a thin layer of water, protecting them from desiccation. The same was observed for the cyprids attached to adult stalks, which were mostly found settled in between the scales. It is unclear whether this attachment pattern at the meniscus is an adaptive characteristic, which in nature may serve to target upper water column levels, or an artefact of laboratory rearing (see Di Fino et al. 2014).

The effects of salinity on settlement have been investigated for several barnacle species (e.g. B. improvisus, B. eburneus, B. subalbidus; Dineen and Hines 1992, 1994a, b). For P. pollicipes, cyprid exploration and attachment decreased (≈10% at 20 psu) at salinities below natural (30-40 psu), in a similar way to metamorphosis rate (≈20%). P. pollicipes inhabit coastal environments where salinity ranges between 33 and 37 psu and it is therefore not surprising that larvae are better adapted to survive and develop within this range. Regulatory mechanisms needed to cope with sub-natural salinities, or the recognition of sub-optimal conditions, might lead to a decrease in surface exploration, but also a reduced capacity to metamorphose. Nevertheless, the tolerance to lower salinities (e.g. 20 psu) is indicative of physiological plasticity, as seen in other barnacle species (e.g. B. amphitrite, Kon-Ya and Miki 1994). This increased tolerance to a range of salinities might relate to their adaptation to the highly-exposed intertidal, where larvae can be subjected to situations of higher salinity (e.g. after water evaporation on rock pools during low tide) or lower salinity on the surface water layer (e.g. after precipitation). Other authors (e.g. Dineen and Hines 1994a, b) have observed that the response to salinity varies significantly with species selectivity, correlating with their distribution range, but can also be heavily influenced by other factors such as age and the use of barnacle extracts, which contain settlement pheromone (Clare 2011).

The mostly positive phototactic behaviour of P. pollicipes nauplii (Molares et al. 2002) also has ecological advantages by placing larvae higher in the water column, thus facilitating their feeding on phytoplankton. In nature, larvae are also subjected to long-day photoperiods during the breeding season. In this study, the overall preference was for permanent attachment in illuminated areas, with rates almost four times higher than those in the shaded areas. Results from other studies indicate that settlement can vary with light, with many species showing positive phototaxis and with species-specific preferences for settlement under direct or diffuse light (e.g. Daniel 1957, Crisp and Ritz 1973, Pawlik 1992), which have been suggested to be adaptive responses. The lack of difference in metamorphosis rate suggests that this process is independent of light.

Temperature has well-documented effects on barnacle larval development and survival (Satuito et al. 1996, Marechal et al. 2004). Results from Coelho (1990) indicate that the use of higher temperatures significantly decreases larval development time in P. pollicipes, as well as survival. This also applies to barnacle cyprids, as higher temperatures increase metabolic rate and oxygen consumption, with effects on the depletion of energy reserves and a reduction in competence to settle (Satuito et al. 1996). Studies on other species (e.g. Thiyagarajan et al. 2003) have shown that settlement can vary according to temperature. Although permanent attachment was not influenced by temperature, metamorphosis rate was higher (≥20%) at higher temperatures (17 and 20°C). This can be related to temperature accelerating development in the first week and not necessarily implying a higher rate of metamorphosis at these temperatures. For example, B. amphitrite larvae reared at 20°C could successfully metamorphose over a shorter period than those reared at 30°C (Anil et al. 2001). Moreover, Pechenik et al. (1993) reported that cyprids that were forced to delay metamorphosis showed slow post-settlement growth, in spite of the lack of effects on completion of metamorphosis and post-settlement survival, highlighting the need for further investigation in P. pollicipes. In nature, temperatures over the breeding season vary between 14 and 20°C from May to October, tending to the upper range from July to September. Faster metamorphosis at higher temperatures might therefore confer a competitive advantage during this period.

Of the cyprid-related factors that might affect settlement on conspecifics in culture, only ageing and larval batch were investigated. Cyprids of Balanus amphitrite become less selective for substrata as they age (Rittschof et al. 1984), consistent with the ‘desperate larva hypothesis’ (Toonen and Pawlik 1994). In the present experiments with P. pollicipes, attachment also increased as larvae aged. Rittschof et al. (1984) also reported that age-dependent responses of cyprids decrease in the presence of settlement inducers, which suggests that if settlement was being tested on substrata other than the adult, these settlement differences could be higher. The temperature at which cyprids are aged is also crucial to this effect. Satuito et al. (1996) found that storage protein (cyprid major protein) in B. amphitrite decreases with time when ageing is carried out at 15, 20 and 25°C, but not at 5°C. However, longer ageing implies decreased storage reserves, which signify reduced competence, or ability to metamorphose, in aged cyprids (Holland and Walker 1975, Lucas et al. 1979, Satuito et al. 1996). This seems to be the case for P. pollicipes in which, after a week, fewer metamorphosed cyprids were observed in stocks aged for 6 days. However, this does not necessarily imply an absolute decrease in metamorphosis, but rather a possible decrease in rate. A settled cyprid will always attempt to moult to a juvenile barnacle (West and Costlow 2005) although a minimum oil cell volume is required to ensure successful metamorphosis. As cyprids age, therefore, the number of cyprids incapable of completing metamorphosis will increase. Furthermore, it should be noted that post-settlement growth can decrease in aged barnacle cyprids, even if metamorphosis rate is not affected (Pechenik et al. 1993). There are also interesting implications with regard to what might be happening in the wild. Barnacle cyprids are known to postpone their settlement for varying periods that differ between species (Knight-Jones 1953, Krug 2006). However, although cyprids of P. pollicipes might be able to survive in the wild for long periods, the ability to successfully attach and metamorphose will likely decrease steadily with time, selecting only the fittest individuals. In the laboratory, S. balanoides lasted 28 days (10°C) on energy reserves, after which the capacity to metamorphose was compromised due to a drop in reserves below the critical minimum for moulting (Lucas et al. 1979). As the quantity of the storage reserves will determine how long settlement can be delayed until a settlement site is identified, the timing of larval release and development rate is of paramount importance.

Batch differences also affected permanent attachment and metamorphosis and could impact the variability of recruitment in the wild. Although in the present study the maximum attachment differences were about 10%, Holm et al. (2000) reported differences of 10% to 40% in settlement for different batches of B. amphitrite larvae, most likely due to parental as well as environmental effects.

The present study presents the first reference values for P. pollicipes larval permanent attachment over time, which follow patterns previously reported for other species (e.g. Rittschof et al. 1984). Metamorphosis was no exception, with approximately 80% of larvae moulting within 12 days and about 10% of larvae failing to metamorphose. Kugele and Yule (1996) reported a similar pattern, with 64% of cyprids metamorphosing in 6 days, 18% failing to metamorphose, and negligible settlement on non-conspecific substrata. The percentage of cyprids failing to metamorphose in the laboratory suggests that the health of laboratory-reared larvae is certainly an issue to consider.

Cyprid survival in the absence of conspecific substrata showed a steady rate of decline to ≈50% after 20 days in culture, reflecting the larval capacity to survive over long periods in the water column. It is further hypothesized that, during summer, the long survival could be prolonged by periods of colder water temperatures associated with upwelling events (Fiuza et al. 1982). Increasing lethargy of larvae over this period might represent an energy-saving strategy to survive for long periods in the wild, when settlement is not possible and larvae are dependent on currents for suitable on-shore transport. However, results for other species indicate particularly long cyprid phases for larvae reared in culture (Batham 1946) and it cannot be excluded that these may be overestimations of survival in the wild.

The results of this study constitute the first laboratory investigation into some of the factors modulating gregarious settlement of this species. Gregariousness occurred over a range of conditions, but attachment and metamorphosis were influenced independently by the various experimental treatments. Larval-related factors affected both attachment and metamorphosis, with cyprid age decreasing both rates. Larvae were nevertheless capable of surviving for extended periods. All abiotic factors tested (except temperature) affected permanent attachment, which was optimal under conditions of light, salinity and hydrodynamics similar to those experienced in nature. On the other hand, metamorphosis was affected by neither light nor hydrodynamism, but was higher at natural salinities and higher temperatures, conditions consistent with the breeding season (Cruz and Araújo 1999) and downwelling periods (Fiuza et al. 1982). Although these observations might suggest that P. pollicipes cyprids could cope well with predicted temperature and salinity fluctuations in the North Atlantic Ocean, further studies should investigate the interaction of these factors, particularly within the context of climate change. Furthermore, although larvae settled heavily on conspecific adults, habitat selection in nature is likely to be mediated by a variety of factors beyond those addressed here, such as transport to the coast, larval choice and post-settlement survival. Further research on the importance of substrata-related factors, both abiotic (e.g. surface topography, desiccation) and biotic (e.g. presence of incumbents, including biofilm, predation, competition and chemical signals/cues), is essential to understand how the interaction of these factors modulates recruitment and constitutes an essential tool for stock management programmes.