Introduction

On the Atlantic coast of North America, heavy encrustations of the non-native epiphytic bryozoan Membranipora membranacea on kelps cause blades to weaken, fragment and subsequently defoliate during autumn storms (Lambert et al. 1992; Scheibling et al. 1999; Saunders and Metaxas 2008). By removing the competitively dominant kelps, outbreaks of M. membranacea have facilitated the introduction and establishment of monospecific stands of the non-native green algae Codium fragile ssp. fragile (Levin et al. 2002; Scheibling and Gagnon 2006). C. fragile is chemically defended against herbivory (Lyons et al. 2007) and its consumption reduces the somatic and gonadic growth of the dominant local grazer, the sea urchin Strongylocentrotus droebachiensis (Lyons and Scheibling 2007). M. membranacea is native to both the Atlantic coast of Europe and the Pacific coast of North America, and was first observed in 1987 in the Gulf of Maine (Berman et al. 1992), and in 1992 on the southern shore of Nova Scotia (Scheibling et al. 1999). In Nova Scotia, larvae of M. membranacea (cyphonautes) begin to settle on kelps in May–July, and reach peak settlement in September–October (Saunders and Metaxas 2007). Upon settlement, larvae metamorphose into a pair of sessile filter-feeding zooids (ancestrula), which then bud asexually to form sheet-like colonies on kelp blades (Fig. 1). After reaching peak coverage on kelps in the autumn (Berman et al. 1992; Saunders and Metaxas 2008), colonies senesce, and <1% of the population overwinters (M. Saunders unpublished data) and provides the larvae for the following year.

Fig. 1
figure 1

Colonies of Membranipora membranacea: (a) newly settled colony (ancestrula) (with the larval shell and 2 zooids present, classified as “0.5 zooid generation”); (b) growing colony with beginning of a new zooid (2.5 zooids present, enumerated as 1.5 zooid generations); (c) colony with ~45 fully formed zooids, enumerated as 7.5 zooid generations [numbers indicate rows from 1 (ancestrula) to 7 (at the growing edge)]; (d) colonies on a kelp blade; (e) tagged colony in the field; and (f) zooids within a large colony—the darker zooids at the bottom were dyed 2 weeks previously with Alizarin Red S, and the lighter portions are new growth. Scale bars indicate: (a, b) 200 μm, (c) 3 mm, (d) 50 mm, (e) 1 cm, (f) 5 mm

Full size image

Outbreaks of M. membranacea resulting in pronounced decreases in kelp canopy cover were observed on the southern shore of Nova Scotia in the autumns of 1993 (Scheibling et al. 1999), 1997, 1999 (Scheibling and Gagnon, unpublished manuscript) and 2006 (Saunders and Metaxas 2008). Outbreaks have been attributed to warmer winters through increases in overwintering survival and earlier and more abundant larval settlement and recruitment (Saunders and Metaxas 2007, 2008). Within a range of thermal tolerance, growth rates of ectothermic organisms are typically positively related to temperature (Trudgill et al. 2005). Consequently, warmer water during the growth period in summer and autumn potentially could result in higher growth rates of colonies and thus contribute to outbreaks. Bryozoan colony growth rates can also be affected by other factors, such as food supply (e.g. O’Dea and Okamura 1999), which could either independently, or in interaction with temperature, contribute to variations in M. membranacea abundance.

Measures of growth rates and size at age are critical for modeling population dynamics and, thus, predicting population outbreaks. For many aquatic organisms, it is logistically difficult or even impossible to measure growth rates in the field. Instead, growth rates are measured in the laboratory and then extrapolated to the wild population. However, laboratory estimates of biological processes may vary considerably compared to field measures. For example, growth and survival rates of Daphnia ambigua were higher in the laboratory than in the field, likely as a result of limited food availability in the field (Lei and Armitage 1980), and consumption rates by three crab species were two to four times higher in the laboratory than in the field (Breen and Metaxas 2008). Moreover, we might expect variations in growth rates to be greater in the field than in the laboratory, due to the many abiotic and biotic variables that cannot be controlled in the natural setting.

Growth rates of M. membranacea have been measured under a variety of conditions in the laboratory (e.g. Menon 1972; Eckman and Duggins 1993), and on artificial substrates in the field (Harvell and Grosberg 1988; Eckman and Duggins 1991; Harvell and Helling 1993; Padilla et al. 1996; Grünbaum 1997; Pratt 2008). However, growth rates of M. membranacea have rarely been quantified on natural substrates in the field (but see Yoshioka 1973; Okamura and Partridge 1999) and survival rates of Membranipora species are lower on artificial compared to natural substrates (Manriquez and Cancino 1996). Within the native range, M. membranacea growth rates are known to increase with temperature (Menon 1972) and size (Yoshioka 1973). However, growth rates within the introduced range in the western North Atlantic, where food supply, temperature, predation, and algal substrates differ from other regions, have not been quantified. Additionally, the combined effects of temperature and size, the interactive effects of temperature and food, and the differences between laboratory and field measurements on the growth rate of M. membranacea are unknown.

We quantified the growth rate [linear extension of colony size (diameter for circular colonies, or maximum dimension for oblong colonies, mm day−1)] of colonies of M. membranacea using a number of different techniques in both field and laboratory studies. We examined the effects of (1) colony size and temperature (field and laboratory), (2) setting (laboratory versus field), and (3) food availability (limited and unlimited, laboratory). We predicted that growth rate would be positively related to size, temperature, food availability for any given temperature, and higher in the laboratory than in the field.

Materials and methods

Field experiments were conducted in St. Margarets Bay and Lunenburg Bay, Nova Scotia (NS), Canada, and off the floating dock at Friday Harbor Laboratories (FHL), Washington (WA), USA. Laboratory experiments were conducted at Dalhousie University, NS, and at FHL. Experiments at FHL were included because: (a) delicate small (<4 mm) colonies could be tagged without the use of SCUBA, since bryozoans on kelps are present on the floating docks at FHL, and (b) the timing and high availability of competent larvae are predictable at FHL. In all experiments, growth rate was measured for colonies that were circular or oblong in shape, and were not crowded by conspecifics or significantly limited by space. Colonies <50 mm (approximately) have a relatively circular shape, whereas those >50 mm become elongated, and maximal growth typically occurs in the direction of the longest axis, for those colonies. Temperature was measured at 10-min intervals using Onset Co. Pendant data loggers (accuracy ±0.47°C), and averaged at daily intervals. In the field, loggers were affixed to the substratum at the same depth and located at 1–20 m from the marked colonies. In the laboratory, loggers were placed in the aquaria or incubators adjacent to the growing colonies.

Laboratory studies

Newly settled colonies, Washington

Settlement-competent M. membranacea larvae (cyphonautes) were isolated from plankton collected with a 333-μm net of the floating dock at FHL. Larvae were maintained in 80-μm filtered seawater at ambient temperature (~12°C) in 100-ml glass dishes for up to 12 h prior to inclusion in experiments. Larvae were introduced into dishes (1 larva per dish) with a single 1 × 1 cm square piece of Laminaria saccharina resting on the bottom to induce settlement. The kelp had been maintained in running seawater for 24 h prior to inclusion to reduce its production of mucilage. The majority of larvae metamorphosed immediately upon introduction to the dishes with kelp, but the dishes were maintained overnight in the seawater tables to maximize the total number of settlers. Colonies on kelps (N = 36) were subsequently introduced into eighteen 100-ml (“food unlimited” treatment) and eighteen 6-ml (“food limited” treatment) individual dishes, which were subsequently placed into three incubators (6 colonies per food treatment per incubator) maintained at each of three temperatures (6, 11, and 14°C; Table 1) with a 14/10-h day/night cycle for 15 days. Seawater was changed on alternate days with: (a) 100 ml of 3-μm filtered seawater supplemented with the microalgae Isochrysis galbana (4 × 104 cells ml−1) and Dunaliela tertiolecta (5 × 103 cells ml−1) (“food unlimited”); or (b) 6 ml of 80-μl filtered seawater (“food limited”). The algal concentration in the “food unlimited” treatment (4.5 × 104 ml−1) was within the range of optimum values reported for various bryozoan species [1.5 × 103–1.0 × 105 ml (reviewed in Amui-Vedel et al. 2007)] and was at the upper end of the range of phytoplankton concentrations observed in coastal Nova Scotia throughout the year (100–104 cells ml−1, Metaxas and Scheibling 1996). Algal cells remained suspended during the 48-h period between water changes. To maximize food limitation, the food limited colonies were grown in a smaller volume of seawater than the food unlimited colonies. Since we observed similar growth patterns for colonies grown in 3-μm filtered seawater in 100 ml containers (with no added food) as for the “food limited” treatments (unpublished data), we do not believe that the container size affected the experimental outcome. Colonies and kelps were cleaned daily by gentle shaking to remove fecal pellets and sedimentation.

Table 1 Details of the field and laboratory experiments to measure growth rate of M. membranacea colonies in Nova Scotia (SMB St. Margarets Bay, Lun Lunenburg Bay, Dal Dalhousie) and Washington (FHL Friday Harbor Laboratories) in 2005–2007
Full size table

On each day, the maximum number of rows of zooids (“generations”) from the ancestrula to the edge of the colony was quantified. For example, settlers which had not yet lost their larval shell (Fig. 1a) were scored as “0.5 generation”, ancestrula (one row of zooids with no larval shell) as “1 generation”, ancestrula plus one incompletely formed zooid as “1.5 generations” (Fig. 1b), and so on (Fig. 1c). We developed a relationship to convert the number of zooid generations in a colony (Z) to the maximum colony diameter in mm (D), by measuring the maximum dimension of colonies over a range of sizes on day 15 in digital photographs (taken with a Q imaging Micropublisher 3.3 RTV camera through a Nikon Eclipse E600 compound microscope) using Sigma Scan Pro image analysis software (D = 0.40e0.36Z, n = 17, R 2 = 0.83, P < 0.001). This technique was employed to reduce handling time each day. Although zooid size in bryozoans can be influenced by temperature (e.g. O’Dea and Okamura 1999; Amui-Vedel et al. 2007), we did not observe such an effect on zooid size in this study (M. Saunders unpublished data), likely because the colonies were small and still within the zone of astogenetic change (Boardman et al. 1970).

Established colonies, Nova Scotia

Large (10–192 mm; Table 1) colonies of M. membranacea growing on the distal ends of the kelp Laminaria digitata were collected using SCUBA from 5 to 7 m depth at Splitnose Point, Nova Scotia (44º28.6′N, 63º32.7′W) in August 2007. The larger colonies had overwintered from the previous year, and were growing in only one direction, the other side having been fragmented off with the distal end of the kelp fronds as the kelps grew. Colonies on kelps were transported to the laboratory in seawater (4°C) in large plastic tubs, and placed in seawater tables at Dalhousie University’s Aquatron facility within 2 h of collection. Colonies were collected for two experiments on 3 August 2007 and 20 August 2007, for Experiments 1 and 2, respectively. Each colony was marked by passing a small numbered cable tie through the kelp at 0.5 mm from the growing edge of the colony. Colonies were assessed daily, and portions of the kelp substrate which were rotten were carefully trimmed off using small scissors (as in Lutaud 1961). Throughout the experiments, colonies that were damaged or were growing on particularly deteriorated kelp were discarded.

Colonies were maintained in three 120-l seawater tables at temperatures of 7, 11, and 14°C for Experiment 1, and 9 and 14°C for Experiment 2, for 13 days (Table 1). In Experiment 1, portions of kelp blades with colonies were maintained in plastic containers [26 × 16 × 18 cm (L × W × H), with 16–18 holes (3 cm diameter) lined with 2-mm mesh to allow water circulation], which were placed in the seawater tables. In Experiment 2, to maximize flow and minimize algal deterioration, kelps were placed directly in the seawater tables. Colonies were fed Tetraselmis chuii (~1 × 106 cells ml−1 culture) daily through a drip system into the flow-through seawater tables. In total, 6.5 and 4.5 l day−1 of culture were added for Experiments 1 and 2, respectively. Flow through each seawater table was approximately 4 l min−1, and thus the volume of seawater in each table was replaced approximately every 30 min. Upon termination of the experiments, colonies were preserved with 7% buffered formalin. Colony growth was measured as the maximum linear distance (mm) from the cable tied at the colony’s edge upon collection, to the edge of the recently grown colony, using a clear plastic ruler (accuracy 1 mm).

Field studies

Tagged colonies, Nova Scotia

Colonies (4–128 mm; Table 1) on kelp (Saccharina longicruris) were tagged in situ using SCUBA at Lodge (44°33.3′N, 64°01.9′W), on the western shore of St. Margarets Bay, Nova Scotia [see Saunders and Metaxas (2007, 2008) for site descriptions]. Colonies were tagged by carefully threading a small plastic cable tie through the kelp at a distance of 1–3 cm from each colony. Colonies were tagged during 3 measurement periods, at 4, 6 and 8 m depth (15–23 August 2005), and at 4, 8, and 12 m depth (12–22 August 2006, and 22 August–9 September 2006). The size (maximum dimension) of the colony (to the nearest mm) was measured with a clear plastic ruler at the time of tagging, and again at the end of the measurement period. Average daily water temperature during the study periods ranged from 8.6 ± 0.8 to 16.2 ± 1.8°C depending on time and depth (Table 1).

Tagged colonies, Washington

Colonies 0.5–4 mm in diameter (Table 1) were tagged on kelps (Laminaria saccharina) at 0.5 m depth on the floating dock at FHL, using small plastic garment tags passed through the kelp ~1 cm from the edge of each colony. To do so, rubber tires on which kelps were growing were lifted from the water for ~20 min. Colonies were tagged and measured (number of zooid generations) on 6, 7 and 12 May 2006, and final colony size was measured on 25 May 2006 (after 19, 18 and 13 days, respectively). Diameter was calculated based on a known relationship between zooid rows and diameter (mm) (Saunders and Metaxas 2008). Average daily water temperature during the study period was 9.9 ± 0.2°C.

Dyed colonies, Nova Scotia

We employed a novel approach for measuring growth rates of M. membranacea in situ. Alizarin Red S is a calcium binding dye that is incorporated into the calcium carbonate structures of actively growing organisms, and has been used extensively to measure growth rates in corals (e.g. Dustan 1975). Colonies (8–66 mm; Table 1) on the kelps S. longicruris and L. digitata were dyed using SCUBA at Sculpin Shoal, Lunenburg Bay, NS (44°20.3′N, 64°15.0′W). The shoal is located near the centre of the bay, and consists of multiple ridges of granitic bedrock, which are oriented parallel to the mouth the bay. Ridges (2 m deep) are separated by valleys of 5–10 m deep and 10–50 m wide. To dye M. membranacea colonies, plastic bags (40 l) were placed over individual kelps, cinched around the stipe, and injected with highly concentrated Alizarin Red S (Sigma Aldrich) in filtered seawater, to achieve an in-bag dye concentration of approximately 100 g ml−1. Bags were closed tightly around the stipe and left for approximately 18 h. The bags were then removed and colonies were allowed to grow. The study was repeated twice (25 July–5 August 2007, and 6–29 August 2007). At the end of the study period, dyed colonies on kelps were collected, returned to the laboratory in coolers (4°C), and preserved in 7% buffered formalin in seawater within 8 h of sampling. Skeletal tissue formed prior to the dye application was stained pink, and any new growth after the removal of dye was white (Fig. 1f), allowing measurement (mm) of both the original colony size and of the new growth.

Data analysis

Growth rate was calculated as the linear extension of colony diameter (for circular colonies) or maximum dimension (for oblong colonies) per day (mm day−1). Using all data except for the growth rates of food limited colonies grown in the laboratory, the effect of setting (field, laboratory) on standardized colony growth rate (day−1) [growth rate (mm day−1)/initial colony size (mm)] was examined using one way ANCOVA with average daily temperature (°C) as the covariate. For each setting independently, the effects of initial colony diameter and temperature on growth rate were examined using multiple linear regressions. The effects of temperature (3 levels: 6, 11, 14°C) and food (2 levels: limited, unlimited) on the size of colonies after 15 days were examined using two-way ANOVA. For the colonies grown with limited food, colony size was observed to increase from days 1 to 6, but not later in the study period. To quantify this apparent trend, we calculated growth rates for 50% of the colonies for days 1–6, and for the other 50% of the colonies from days 6 to 15, for each food treatment level (n = 3 per treatment per time period). The effects of time period (2 levels: days 1–6, days 6–15), temperature (3 levels: 6, 11, 14) and food (2 levels: limited, unlimited) on growth rate were examined using three-way ANOVA. Colony diameters were log-transformed for the ANCOVA and for the regressions examining the effects of setting, size, and temperature on growth rate. Colony growth and standardized growth rates were log-transformed for all analyses, except for the analysis of the effect of food treatment, where growth rates were log(x + 0.1)-transformed. The data satisfied the assumption of homogeneity of variances according to Levene’s test, but did not satisfy the assumption of normality according to the Shapiro–Wilk test; however, ANOVA is robust to deviate from normality (Zar 1999). The data for the effects of food × temperature × time period on growth rate did not satisfy the assumption of homogeneity of variances according to Levene’s test because there was no variability for 2 of the treatments (10°C/No Food/days1–6, and 14°C/Food/days1–6), thus inflating the probability of a type I error (Zar 1999). Consequently, we used a more conservative αcrit(0.01). When appropriate, homogeneous subsets were identified using Student–Newman–Keuls test. Statistical analyses were conducted using SPSS 15.0.

Results

Effects of size, temperature and setting on growth of colonies

Colony growth rates of M. membranacea ranged from 0.01 to 12 mm day−1. For any particular size, growth rate spanned almost one order of magnitude, depending on setting and temperature. For example, colonies 0.5–0.6 mm in diameter grew 0.01–0.1 mm day−1, and those 100 mm in diameter grew 1–10 mm day−1. Standardized colony growth rate increased with temperature and was significantly higher in the field than in the laboratory (Table 2; Fig. 2). The relationship between growth rate (GR, mm day−1), initial colony size (S, mm) and temperature (T, °C) was described by the equations:

$$ {\text{Laboratory}}:\quad { \log }\left( {\text{GR}} \right) = - 1.683 + 0.677\cdot{ \log }\left( S \right) + 0.058\left( T \right)\quad \quad R^{2} = 0.87,{\kern 1pt} \;P < 0.001 $$
$$ {\text{Field: }} {\log }\left( {\text{GR}} \right) = - 1. 6 6 5+ 0. 7 1 9\cdot{ \log }\left( S \right) + 0.0 7 2\left( T \right)\quad \quad \quad R^{ 2} = 0. 8 4,\;P < 0.00 1 $$

The standard errors of the regression constant, S parameter and T parameter were 0.137, 0.036, 0.012 (laboratory), and 0.086, 0.032, and 0.008 (field), respectively. In both settings, size explained more of the variability in growth rate than temperature (partial correlation: field: S, r = 0.85, and T, r = 0.56; laboratory: S, r = 0.93, and T, r = 0.55).

Table 2 Results of ANCOVA examining the effects of setting (laboratory, field) and temperature (covariate) on the log-transformed standardized growth rates (growth rate/initial colony size) of Membranipora membranacea colonies grown in Nova Scotia, Canada and Washington, USA
Full size table
Fig. 2
figure 2

Standardized growth rate (growth rate/initial size) of colonies of Membranipora membranacea measured in the field and the laboratory over a range of sizes and temperatures. NS Nova Scotia; SMB St. Margarets Bay, NS; Lun Lunenburg Bay, NS; FHL Friday Harbor Laboratories, Washington. Field and laboratory data are indicated in white and gray, respectively

Full size image

Effects of food and temperature on growth of newly settled colonies

At both food levels, newly settled colonies lost their larval shell after 2 days at 14°C, 2–3 days at 11°C, and 5 days at 6°C. There was a significant 2-way interaction for the effects of food and temperature on the size of colonies at the end of the study period (Table 3). The largest colonies (2 mm) were those grown at 14°C under unlimited food (Fig. 3). Those colonies grown at 11°C under unlimited food were of intermediate size (1.5 mm), and the sizes of all other colonies (unlimited food at 6°C, and limited food at 6, 11, and 14°C) were smallest (0.8–0.9 mm) (SNK tests).

Table 3 Results of two-way ANOVA examining the effects of temperature and food on the diameter of newly settled colonies of Membranipora membranacea after 15 days of growth in the laboratory
Full size table
Fig. 3
figure 3

Diameter (mean ± SE, N = 3) of colonies of Membranipora membranacea grown from newly settled larvae for 15 days in the laboratory at 3 temperatures with 2 food treatments

Full size image

There were significant 3-way and 2-way interactions of the effects of food, temperature, and time period on colony growth rate (Table 4; Fig. 4). Colony growth rate varied, only for the food limited colonies, between time periods at 11 and 14°C but not at 6°C. For days 1–6, growth was fastest in the 14°C water and slowest in the 6°C water for both food levels, and there was no effect of food at any temperature. For days 6–15, growth rate was higher for colonies at 11 and 14°C with unlimited food than the growth in all other treatments. At 6°C, there was no effect of food or time period on growth rate; however, at 11 and 14°C, growth during days 6–15 was slower when colonies were grown with limited food than for the other treatments.

Table 4 Results of three-way ANOVA examining the effects of temperature, food and time period on the growth rates of newly settled colonies of Membranipora membranacea in the laboratory
Full size table
Fig. 4
figure 4

Growth rate (mean + SE, n = 3) of newly settled colonies of Membranipora membranacea grown during 2 time periods (days 1–6 and days 6–15) in 3 temperatures, and with 2 food treatments. Letters above bars indicate homogeneous subsets identified using Student–Newman–Keuls test

Full size image

Discussion

The combination of colony size and temperature explained 84–87% of the variability in growth rate of relatively uncrowded M. membranacea colonies during summer. We observed increasing growth rates with respect to temperature between 5.7 and 16.2°C, which is within the range of temperatures observed in St. Margarets Bay, NS [<0–19°C (Saunders and Metaxas 2007)]. Menon (1972) found a similar result for colonies from the North Sea grown in the laboratory at 6–18°C, and reported the lower lethal temperature as <0°C for colonies acclimated at 6 or 12°C, and 1.5°C for colonies acclimated at 18°C. Because of the strong dependence of growth on temperature, changes in temperature either in winter/spring, when a few remnant colonies are overwintering, or in summer/autumn, when the population is growing, could have a significant impact on the abundance of this non-native species. However, the relative importance of variations in temperature at particular times of year on outbreak dynamics of M. membranacea have not been examined. An interaction between temperature variability and the population dynamics of introduced species has been shown for other organisms. In New England, 6 species of non-indigenous and indigenous tunicates had similar growth rates at moderate temperatures, but the non-indigenous species had more pronounced increases in growth rate with temperature than the native species (Stachowicz et al. 2002). In contrast, in Santa Barbara harbor, recruitment and survival of the non-native kelp Undaria pinnafida was linked to periods of cooler water (Thornber et al. 2004). The specific relationships between temperature and growth of introduced species are likely determined by the climate of origin of the species.

There were no significant effects of food on the growth of newly settled colonies at any temperature for the first 6 days, suggesting that there was enough energy from larval reserves and in the 6 ml of 80-μm filtered seawater to sustain growth until a diameter of 0.8 mm was reached. Growth rates of newly settled colonies during the second week of growth with limited food were not affected by temperature, in contrast to colonies grown in unlimited food conditions. In the northwest Atlantic, phytoplankton abundance is typically low in winter, and high in spring and summer (Townsend and Thomas 2001; Yoder et al. 2002), with peaks during the spring and fall blooms. However, event-scale (days–weeks) processes such as upwelling can cause significant changes in phytoplankton species composition and abundance (Côté and Platt 1983), which could in turn, affect the growth rates of filter-feeding species such as M. membranacea on short time scales. Each of our field measurements in Nova Scotia integrated growth rate over periods of ~2 weeks, and were obtained from July to September during 3 years; thus, they incorporate variability in growth due to variations in food during times of relatively abundant food. The relationship between temperature- and size-dependent growth rates may differ during other times of year (e.g., winter) when colonies are senescent or food is particularly scarce. Standardized growth rate was very low at 6°C, and at such temperatures food did not affect growth of newly settled colonies, similarly to established colonies of the bryozoan Conopeum seurati in the field (O’Dea and Okamura 1999). This suggests that in Nova Scotia, low food abundance in winter is of much lower importance to colony growth than the extremely cold (−1 to 3°C) temperatures.

The growth rates we report here (up to 12 mm day−1) are higher than those for many other bryozoan species [e.g., Conopeum seurati: <1 mm day−1 (O’Dea and Okamura 1999), Pentapora foliacea: 2 cm year−1 (Pätzold et al. 1987)]. Although r (intrinsic rate of increase) of this species is of intermediate value relative to other bryozoan species (Hermansen et al. 2001; Amui-Vedel et al. 2007), the growth rates of many other bryozoan species decelerate at a particular size (Cancino and Hughes 1987) or are inversely related to size (Cocito et al. 1998). In contrast, we observed a positive and exponential relationship between size and growth for all colonies within the size range examined (0.5–192 mm). In Nova Scotia, colonies on Saccharina longicruris can reach maximum dimensions of almost 400 mm (Saunders and Metaxas 2008), which is 1–2 orders of magnitude larger than typically observed for the other local epiphytic invertebrates (Electra pilosa, Obelia geniculata, Spirorbis spirorbis). In field and laboratory studies, M. membranacea had faster growth rates, reached larger sizes, and had higher survival rates than three other bryozoan species (Pratt 2008); as a result, M. membranacea is competitively superior to the other epiphytic species on kelps (Berman et al. 1992). In combination with a lack of predators, high growth rates and the ability to reach large sizes contribute to M. membranacea’s success as an invasive species in the western North Atlantic.

A number of factors other than those we examined can affect growth rates of bryozoans, and may explain the remaining 15% of the variation in our study not accounted for by temperature and colony size. Growth rate of reproductive colonies can be lower than that of non-reproductive colonies (Harvell and Grosberg 1988), and there are energetic trade offs between growth and defense (tower cells, spines, stolons) (Padilla et al. 1996; Grünbaum 1997). Flow velocity has an unclear effect on M. membranacea growth rate, with a negative relationship reported for laboratory (Eckman and Duggins 1993) and field (Pratt 2008) studies in Washington, but with no effect in a field study in Ireland (Okamura and Partridge 1999). When colonies encounter a conspecific or if they reach the edge of a kelp blade, they will cease growth in that direction (Harvell et al. 1990). Colonies may be partially removed when the underlying kelp tissue breaks, and damage in part of a colony causes increased growth in the direction of the surviving edge (Harvell and Helling 1993). Furthermore, colonies can senesce and shrink, a process that tends to occur in the autumn (Harvell et al. 1990). Thus, temperature- and size-dependent growth rates will likely decrease in the fall as colonies become space limited and begin to senesce. Nevertheless, the growth rates reported in this study are based on a wide range of individuals, substrates, times, and geographic locations, and thus incorporate variability generated by many extraneous factors. Consequently, the strong relationships that we observed appear to dominate over other potential sources of variability.

For logistical purposes, we measured the growth rates of very small (0.5–4 mm) colonies in the field and in the laboratory in Washington. It is possible that the growth rates of colonies in this size range are slightly different in Nova Scotia than in Washington. However, it is unlikely that this difference between sites would be significant compared to the 4 orders of magnitude variation in growth we recorded. Furthermore, although temperature data were not available, size-dependent growth rates of M. membranacea measured in the field in California (Yoshioka 1973) were within the range as we report. Thus, we believe that a coast-specific temperature- and size-specific growth relationship is unlikely for the size range that we examined.

We determined that over a large size range, the growth rates of Membranipora membranacea colonies were significantly higher in the field than in the laboratory. Therefore, for the purposes of generating predictive models of M. membranacea outbreak dynamics, field estimates of growth rates are preferable to laboratory estimates. For small colonies, measuring growth rates in the laboratory allowed for more controlled conditions, less physical disruption to the colonies, and more frequent measurement opportunities. It was preferable to measure growth rates of large colonies in the field, however, because transport to the laboratory was more physically disruptive than in situ tagging. Moreover, large colonies were difficult to maintain in the laboratory (Lutaud 1961) for more than 1–2 weeks because of deterioration of their host kelps, which worsened with increasing temperature. An alternative approach could be to run experiments in the laboratory, using colonies on artificial substrates (e.g. Pratt 2008). However, in both the laboratory and the field, survival of M. isabelleana was lower on artificial substrates than on algae (Manriquez and Cancino 1996). In the field, the tagging method required only one day for marking, but was not useful when colonies were densely aggregated on a single kelp blade because of the difficulty in identifying the colony which had been tagged, or for exposed locations, because wave action would cause the tags to be lost or to abrade the kelps and colonies. The dye method required 2 days in the field, but was more effective for marking colonies of irregular shapes, because it left a permanent trace of the entire outline of the colony on the day it was dyed. This method is also not recommended for wave exposed locations because when applied in shallower (2 m) water, wave action caused the dye-filled bag to damage the growing edge of the colonies (unpublished data). Otherwise, the dye did not appear to adversely affect the growth of the colonies. When the dyed colonies were preserved in formalin, the dye deteriorated with time; however, the zooids had a noticeably different shape at the location where the dye band had been located, permitting analysis of the samples even after extended (months–years) durations of storage.

In summary, we quantified size- and temperature-dependent growth rates of the bryozoan Membranipora membranacea, demonstrated that colonies grew significantly faster in the field than in the laboratory, and found an interactive effect between food availability and temperature. The reported relationships between growth rate, temperature and colony diameter are necessary for developing predictive models of M. membranacea outbreaks, and thus understanding the population dynamics of this ecologically damaging species within the introduced range.