Larval recruitment of the invasive colonial ascidian Didemnum vexillum , seasonal water temperatures in New England coastal and offshore waters , and implications for spread of the species

The invasive colonial ascidian Didemnum vexillum occurs in Japan, North America, northern Europe, and New Zealand. It forms adhering mats on living and non-living hard substrates and alters seabed habitats. We examined the relationship of the first and last occurrences of D. vexillum recruits to water temperature and suggest it is possible to identify coastal and offshore locations that are at risk of colonization by the species based on seasonal water temperature trends. Temperature loggers and settlement plates for recruits were deployed and monitored at three shallow coastal settings in New England, including an open harbor, a marine estuary, and a nearshore island. In addition, the distribution of D. vexillum at sites on Georges Bank, an offshore fishing ground (where settlement plates could not be deployed), was compared to long-term bottom temperature data. Recruits of D. vexillum are small (0.2 to 1.0 mm) but easy to identify, and photographs illustrating the developmental stages of oozooids are presented. Recruits of D. vexillum did not appear on settlement plates at the same water temperature at the three coastal sites. Recruitment occurs in the range of 14 to 20°C and apparently is dependent on local climatic conditions. At coastal sites where we have complete records, recruitment persisted for 3.5 to 5 months; and recruits continued to appear as waters cooled to below the temperature of first occurrence. Recruitment ceased in the range of 9 to 11°C. On Georges Bank, the yearly temperature range (4 to 16-17°C) is the same in areas where D. vexillum is present and in a nearby area where D. vexillum is absent. However, in the reproductive season on the bank, temperature variability is but a few degrees in areas where the species is present; whereas it is high (11°C) where the species is absent. It is known from previous studies that in extremely cold conditions in shallow water D. vexillum colonies degenerate and all but disappear, only to regenerate as waters warm. We suggest that: 1) the degree to which colonies degrade in the cool season influences the length of time they require to regenerate, reproduce sexually, and brood larvae; 2) larval recruits will be released at the end of a developmental period as water temperatures warm, not necessarily when a particular water temperature is reached; 3) larvae likely will appear at different temperatures at climatically different sites, and at approximately the same time and temperature at climatically similar sites; 4) highly variable temperatures during the warm season likely inhibit the reproductive process; 5) the time required for a colony to develop and release larvae and the length of the warm season probably affects the length of the recruiting period at a site; 6) as waters cool, larval release by healthy colonies and recruitment might be regulated chiefly by declining temperature and therefore could end at approximately the same temperature at all sites. At deeper water sites, where minimum temperatures are warmer than at shallow sites, it is possible that D. vexillum colonies are not as affected in the cool season (i.e., do not degenerate) and thus could have a longer recruiting season. This may explain the species’ successful colonization of several large areas of gravel habitat on Georges Bank.


Introduction
Didemnum vexillum Kott, 2002 is an invasive colonial ascidian that attaches to hard living and non-living substrates.The species forms mats on seabeds and other surfaces where currents are strong, and it hangs in clumps from surfaces such as floating docks where currents are weak.It overgrows other ascidians, mussels, oysters, bryozoans, hydroids, sponges, anemones, seaweeds, and barnacles (among others), alters habitats, and fouls harbor facilities, vessel hulls, and aquaculture equipment.Didemnum vexillum reproduces asexually by budding to expand colony size and sexually to produce larvae to colonize new sites.Larvae are brooded within the tunic and when released swim for a period of hours to find and attach to a suitable substrate.Fragments of colonies can reattach and grow, and serve to spread the species to new locations.The species, like most colonial ascidians, does not thrive in muddy habitats or where sediment is mobile, as colonies can be smothered and their oral siphons clogged with sediment particles.The species can tolerate temperatures that range from a low of -1 to -2°C to a high of at least 24 to 25°C.Didemnum vexillum now occurs in temperate waters in many parts of the world, including: Japan; northern Europe (France, Netherlands, Ireland); the U.S. east coast (from Long Island, NY to Eastport, ME near the border with Canada); the U.S. and Canadian west coasts (California, Washington, British Columbia); and New Zealand (North and South Islands).For further background and source materials for this description of D. vexillum see: Coutts (2002), Bullard and Whitlatch (2004), Bullard et al. (2007), Coutts and Forrest (2007), Minchin (2007), Valentine et al. (2007a, b), Whitlatch and Bullard (2007), U.S. Geological Survey (2008), Woods Hole Oceanographic Institution (2008), Lengyel et al. 2009 (this issue).
Recent morphologic and genetic studies of species of Didemnum collected worldwide have shown that the valid taxonomic name for "Didemnum sp." and "Didemnum sp.A" of recent reports is Didemnum vexillum Kott (2002), described from New Zealand; and that Didemnum vestum Kott (2004), described from New Hampshire, is a junior synonym of D. vexillum (Lambert, this issue, Stefaniak et al., this issue).We follow this taxonomy here.
The importance of water temperature in the timing of sexual reproduction in marine animals was described in the classic paper by Orton (1920).Several recent reviews summarize the role of temperature as an important environmental factor in ascidian reproductive processes (Bates 2005;Lambert 2005a, b).Little is known of the role of temperature in the release of larvae in the genus Didemnum.Millar (1958) reported on the effects of temperature on the breeding season of Didemnum candidum in Scottish waters by plotting the presence of developing embryos and brooded larvae against temperature.
He showed that brooded larvae were most abundant in late June, before the warmest part of the year.
We focused on recording first and last occurrences of D. vexillum recruits, and our methods of settlement plate deployment, maintenance, and observation reflect that approach.As we did not inspect the plates every day, the timing and temperature of first and last occurrences is influenced by the number of days between plate inspections.Further study is required to document the day-by-day development of D. vexillum recruits so that the age of their developmental stages can be more accurately estimated.
At three coastal sites in New England (Great Harbor, Appledore Island, Damariscotta River), we used recruitment data from settlement plates and related temperature data.We observed the relative abundance of recruits, but we did not count them.Offshore on Georges Bank, we used temperature data but could not use recruitment plates and so relied on our collected samples and knowledge of the distribution of D. vexillum on gravel habitats (Valentine et al. 2007b).The purpose of the present study is to determine the relationship of seasonal water temperatures to the occurrence of recruits of D. vexillum.This information can be used to identify sites potentially vulnerable to colonization by the species and to develop strategies for managing environments that have been adversely affected or threatened.

Study sites
Great Harbor, Woods Hole, Massachusetts, 2007 The study site (Figure 1) is located in a large open area of the Iselin Dock complex at the Woods Hole Oceanographic Institution (41.524°N, 70.672° W).The area measures 6 x 24 meters and is used for conducting biological studies and to test equipment.Water depth is approximately 16 m.It is bounded on 3 sides by pilings of cement and steel; on two of these sides wood sheathing extends to about 1.2 m below mean lower low water (MLLW).The fourth side of the area is open to the sea.The dock pilings and sheathing are colonized by attached epifauna, including mussels, barnacles, bryozoans, tunicates, sponges, and anemones, among others.Didemnum vexillum is the dominant tunicate and is common to a depth of approximately 5 m below MLLW.
Appledore Island, Isles of Shoals, Maine, 2007 The study site (Figure 1) is located off the western shore of Appledore Island near the Shoals Marine Laboratory of Cornell University and the University of New Hampshire (42.989°N, 70.616°W).Submerged wooden supports of a former dock lie approximately 10 m off the rocky shore to a water depth of approximately 7 m and are dominated by colonial tunicates and fewer sponges, bryozoans, solitary tunicates, and hydroids.Didemnum vexillum is the dominant tunicate and occurs below the intertidal to a depth of 7 m at this site.

Damariscotta River, Maine, 2006
The study site (Figure 1) is located at a pier off Wentworth Point in the Damariscotta River marine estuary at the Darling Marine Center of the University of Maine (43.935°N, 69.581°W).Water depth is approximately 10 m.A floating dock at the end of the pier and adjacent pier pilings are colonized by sponges, bryozoans, hydroids, and solitary and colonial ascidian species, which are the dominant members of the community.Didemnum vexillum is the dominant tunicate and occurs to a depth of 6 m.
Georges Bank, off New England, 2002-2007 Study areas are located at water depths of 45 to 65 m on the northern part of Georges Bank, an important offshore fishing ground that lies in both U.S. and Canadian waters (Valentine et al 2007a).Two adjacent areas of pebble gravel habitat in U.S. waters totaling 230 sq km are colonized by D. vexillum (Figure 1).One area, Ga (centered at 41. 90°N, 67.45°W), is open to fishing, and the other, Gb (centered at 41.95°N, 67.31°W), lies 9 km to the east in a zone closed to fishing near the U.S./Canada boundary.Tunicate colonies cover 75 percent of the pebble gravel at some sites based on video and photographic surveys conducted from 2002 to 2007.Surveys conducted during the same period have identified another area of gravel habitat, Gc (centered at 42.10°N, 67.25°W) 4 to 6 km north of the colonized areas where D. vexillum does not occur (Figure 1).The western part of Gc is open to fishing.

Settlement plates
Great Harbor, Woods Hole, Massachusetts, 2007 Horizontal settlement plates made of medium gray polyvinyl chloride (PVC) 12 x 12 x 0.6 cm were hung on 2 rope strings.Each plate had a central hole for the rope, with a slot extending from the hole to one edge so that plates could be removed for inspection.On each string, fourteen plates were spaced approximately 0.3 m apart from MLLW to 4.5 m, for a total of 28 plates.The 2 rope strings were hung 6.5 m apart alongside the wood sheathing of the dock which extends to 1.2 m (MLLW) below which the plates were within 1 m of pilings to a depth of 4.5 m (MLLW).Both the wood sheathing and the pilings are colonized by D. vexillum colonies, and plates were hung at places where colonies were most common.
During the inspection process, plates were removed from the string, held in a water bath, inspected visually in air with a magnifying lens, and photographed while covered by water.Both upper and lower plate surfaces were inspected.Care was taken to keep the plates wet during the inspection process.The occurrence of D. vexillum recruits was recorded at the site and later confirmed from analysis of the digital images.After each inspection, some or all attached organisms, including small colonies of D. vexillum, were removed to provide space for new recruits.Settlement plates were deployed from April 10 to December 18, 2007 and were inspected 44 times during this period (Table 1).

Appledore Island, Isles of Shoals, Maine, 2007
Vertical settlement plates made of Plexiglas 10 x 10 x 0.6 cm were attached to wire grates that were fixed to submerged horizontal wooden support beams on the southwest face of a former pier.Four plates were deployed 2.0 to 2.5 m below low tide.For analysis, plates were replaced with clean ones and those removed were maintained in a water bath and taken to the laboratory where D. vexillum recruits were identified under a dissecting microscope.Settlement plates were deployed from June 7 to July 30, 2007 and were inspected every two weeks during this period (Table 2).Horizontal settlement plates made of Plexiglas 10 x 10 x 0.6 cm were hung from a floating dock at the end of a pier.Four to six plates were deployed at 0.5 to 1.0 m below the water surface, with the recruitment surface facing the seabed.The plates were analysed in the same manner as at the Appledore Island site.Settlement plates were deployed from April 29 to December 10, 2006.The plates to be retrieved on November 12 were lost.Plates were inspected every two weeks during this period, except on three occasions when the intervening period was three to four weeks (Table 3).

Georges Bank, off New England
It is not possible to conduct settlement plate experiments on the Georges Bank fishing ground.However, sampling over several years during the warm season (August 2006, November 2004) in study areas Ga and Gb has confirmed the presence of mature brooded larvae in D. vexillum colonies.

Water temperature records
At Great Harbor in Woods Hole, water temperature loggers were deployed at 1, 2, 4, 5, 6, 8, and 10 m below mean lower low water (MLLW) and recorded seasonal temperatures at 10-minute intervals to an accuracy of 0.1°C from January 1 to December 18, 2007 (Figure 2).At the Appledore Island site, a temperature logger was deployed at 2.0 to 2.5 m below low tide and recorded temperatures at 60-minute intervals to an accuracy of 0.2°C from June 7 to July 30, 2007 (Figure 3).This is a short record, so data from a nearby buoy were used to characterize the  4 seasonal temperatures at the site (see below).At the Damariscotta River estuary site, a temperature logger was deployed at 0.5 to 1.0 m below the water surface and recorded temperatures at 60-minute intervals to an accuracy of 0.2°C from January 1 to December 12, 2006 (Figure 4).The logger failed to record data from July 6 to July 29, 2006.However, the seasonal temperature trend displays little variability and allows temperatures to be extrapolated with confidence for the missing days.
For comparison with earlier years, temperature data were acquired for all three sites from other sources (Table 4).Great Harbor temperature data for 2001 to 2006 were provided by the National Oceanographic and Atmospheric Administration (NOAA) tide gauge station (8447930) located at our study site (NOAA, 2008).Damariscotta River temperature data for 1997 to 2000 were acquired at our study site by the Darling Marine Center (DMC, 2008).At Appledore Island there is no long term temperature record, so comparative data (from 1 m water depth) were obtained for 2002-2004 from the University of New Hampshire (UNH) environmental monitoring buoy at the Open Ocean Aquaculture site located 5.1 km due south of Appledore Island (UNH, 2008).The most   In the Georges Bank study areas, it was not possible to deploy continuous temperature loggers, and annual bottom temperature trends Note the low variability of temperature in the warm season (JD 210-300) and compare with Figure 7 were derived from individual readings collected during all seasons from 1977 to 2007 on scientific cruises of the National Marine Fisheries Service (NOAA, 2007).In order to characterize the seasonal bottom temperatures of the two adjacent areas (Ga, Gb) colonized by D. vexillum (Figure 1), all data points from an area enclosing the two affected areas were plotted to give a yearly temperature graph compiled from data collected over a 31-year period (Figure 6).A similar approach was used for area Gc (Figure 7) that lies 4 to 6 km to the north of the colonized areas, where D. vexillum has not been observed despite multiple surveys for benthic invertebrates from 1994 to 2007 (Valentine et al. 2007b).

Recognizing recruits of Didemnum vexillum
The brooded tadpole larvae of species of the genus Didemnum are thought to swim at most for several hours (Berrill 1950;Olson 1985) before settling onto a hard substrate to begin metamorphosis into the initial zooid (oozooid) of a colony.At Great Harbor in Woods Hole, we have photographically documented the development of oozooids, which range in length from 0.2 to 1.0 mm.They are smaller than recruits of the colonial ascidian Botrylloides violaceus with which they often co-occur.In very young oozooids of D. vexillum, paired lateral organs of the thorax (LOTs) appear crescent-shaped and are a pale yellow-brown color; black "eyespots" represent the ocellus and statolith (Figures 8, 9).As the oozooid develops, the LOTs become white (Figure 10) because they produce the calcium carbonate spicules which later migrate throughout the tunic (Kniprath and Lafargue 1980).With continued development of the oozooid, spicules outline the oral siphon and cloacal aperture (Figures 10,11).The oral siphon is located at the anterior end of the oozooid, just forward of the LOTs, and the cloacal aperture is located dorsally, between the posterior parts of the two LOTs.The digestive  organs, yellow-brown in color, are located in the posterior part of the oozooid; and feeding by the oozooid produces brown, spherical fecal pellets in the gut area (Figure 10).As growth proceeds, white spicules accumulate in the tunic, obscuring internal organs, and the oozooid forms a colony by asexual budding (Figures 12,13).The most distinctive features of the D. vexillum oozooid are the paired, white crescent-shaped LOTs which are easily distinguished with a magnifying lens (Figure 11).We have not yet determined the time period required for settled tadpole larva of D. vexillum to metamorphose into an oozooid and for the oozooid to bud into a colony, but see Bullard and Whitlatch (2004).

Occurrence of Didemnum vexillum recruits and associated water temperatures
Great Harbor, Woods Hole, Massachusetts, 2007 Water temperatures were recorded at 7 depths to 10 m (MLLW) at the Woods Hole site from January 1 to December 18, 2007.Analysis of the average daily temperature records at all depths showed that the water column to 10 m was well mixed, and that all settlement plates (MLLW to 4.5 m) experienced a similar temperature regime.The record at 2 m water depth is used for illustration (Figure 2).Plates were deployed from April 10 to December 18, 2007 (Table 1), a period during which the initial water temperature was 4.9 o C, rose to a maximum of 23.8°C on August 4, and declined to 3.4°C on December 18.The first recruits of D. vexillum were observed on June 26, 2007 on the bottom sides of plates at depths of 0.5 to 2.8 m at the end of a 4-day interval during which the water temperature increased from 18.7 to 19.4°C (Figure 2, Table 1).Throughout the experiment, the bottoms of plates attracted by far the most recruits, with larvae settling on plates at all depths to the deepest plate at 4.5 m.Although no attempt was made to count recruits, visual observation indicated they settled in highest numbers in July and August and decreased in numbers in September, October, and November.
The last recruits were observed on November 27, 2007, at the end of a 8-day interval during which water temperature decreased from 9.8 to 9.3°C (Table 1).Thus, recruiting at Great Harbor persisted from June 26 to November 27, a period of 5 months.These last recruits were monitored on 4 occasions over 21 days until December 18 when the water temperature had cooled to 3.4°C.During this period, the recruits retained eyespots, and displayed LOTs, increasing numbers of spicules in the tunic, and fecal pellets, indicating they were feeding and developing, but they did not bud new zooids to form colonies, a possible effect of low water temperatures.
Sampling of large, established D. vexillum colonies at the study site on January 4, 2008 (at 2.2°C) revealed that they contained welldeveloped brooded, unhatched larvae that displayed eye spots, tails, and adhesive papillae (used to attach larvae to a substrate).These colonies exhibited brown spots in the tunic and their cloacal canals were clogged with fecal pellets, an indication that zooids were no longer feeding properly and that the colonies likely were beginning to degenerate in response to cold conditions (Valentine et al. 2007a, Figures 10, 11).

Appledore Island, Isles of Shoals, Maine, 2007
Water temperatures were recorded at 2.0 to 2.5 m below low tide from June 7 to July 30, 2007 (Figure 3).Plates were deployed during this period, and water temperatures ranged from 10.5 to 19.2°C.The first recruits of D. vexillum were observed on July 17, 2007, at the end of a 13-day interval during which water temperatures increased from 14.1 to 15.1°C (Figure 3, Table 2).Recruits settled until July 30 when the plates were removed.The recruitment record here is relatively short, and the last appearance of recruits was not recorded.

Damariscotta River, Maine, 2006
Water temperatures were recorded at 0.5 to 1.0 m below the water surface from January 1 to December 12, 2006 (Figure 4).Plates were deployed from April 29 to December 10, a period during which the initial water temperature was 8.4°C, rose to at least 20.0°C on July 30, and declined to 6.8°C on December 10.Recruits of D. vexillum were first observed on July 14, 2006 (Table 3), at the end of a 16-day interval during which the water temperature increased from 15.8 to 18.0°C (extrapolated from data shown in the graph in Figure 4).Recruits were most abundant in late July, August, and early September, and declined in number in October.The last new recruits were observed on October 28 at a temperature of 11.3°C, but we are uncertain if this represents the actual last occurrence because the plates to be retrieved on November 12 were lost.Although there is a gap in the recruitment data for November 12, we know that recruits did not occur below a temperature of 10.0°C, because no recruits were observed on November 28 (Table 3).Recruitment at this site lasted at least 3.5 months (July 14 to October 28, 2006), and possibly for as long as 4 months.

Characteristics of seasonal water temperature cycles in coastal and offshore waters
In Great Harbor from 2001 to 2007, yearly minimum temperatures ranged from -2.5 to 0.6°C, excepting the abnormally warm winter of 2002, and the yearly maximum temperatures ranged from 22.5 to 24.1°C (Table 4).The Damariscotta River estuary site lies 160 km in latitudinal distance to the north of Great Harbor.The yearly minimum temperatures from 1997 to 2000 and in 2006 ranged from -0.1 to 1.4°C, and yearly maximum temperatures ranged from 17.1 to 20°C (Table 4).The annual temperature range is somewhat narrower here than at Great Harbor, with Damariscotta River having milder cool and warm seasons.Appledore Island is located between the Damariscotta River and Great Harbor sites.It lies 10 km off the coast of Maine and is surrounded by Gulf of Maine waters.At the buoy located 5.1 km south of Appledore Island, yearly minimum temperatures from 2002 to 2004 ranged from 1.5 to 4.2°C and are appreciably warmer than at either Great Harbor or the Damariscotta River (Table 4).Yearly maximum temperatures during these years at the buoy site ranged from 18.2 to 20.4°C.The record we collected at Appledore Island in 2007 is incomplete, but the yearly maximum temperature recorded was at least 19.2°C.
The Georges Bank study areas (Ga, Gb, Gc) are pebble gravel habitats that lie far off the New England coast on a relatively shallow bank, and at much greater water depths (45 to 65 m) than the coastal sites (Figure 1).Plots of the long term seasonal temperature trends in areas colonized by D. vexillum (Ga, Gb), and in the uncolonized area 4-6 km to the north (Gc), show the yearly minimum is 4°C or slightly higher, and the yearly maximum is at most 17°C (Figures 6, 7).The chief difference between the colonized and uncolonized areas on the bank is the variability of bottom temperatures during the warm season (JD 210-300).In areas colonized by D. vexillum (Ga, Gb), temperatures vary only 4°C (from 13 to 17°C); whereas in the area free of the species (Gc), temperatures vary 11°C (from 5 to 16°C).

Didemnum vexillum recruits and recruitment patterns
Little is known of the behavior of the tadpole larvae of D. vexillum.We assume their behavior is similar to tadpoles of other species of Didemnum, which are short-lived and settle within hours of leaving the colony.Berrill (1950) reported that the free-swimming period of Didemnum gelatinosum lasted approximately 2 hours.Olson (1985) showed that tadpoles of Didemnum molle usually settled in 10-15 min but could swim for up to 2 hours.We placed our settlement plates adjacent to substrates heavily colonized by D. vexillum to increase the probability that many larvae would be near the plates and to minimize the time between release and settlement of larvae.In our study, both vertical and horizontal plates successfully attracted larvae, and the lower surfaces of horizontal plates were favored over upper surfaces.
Didemnum vexillum recruits first appeared at different water temperatures at the three coastal study sites.At Great Harbor, Damariscotta River, and Appledore Island, first recruits were recorded in the range of 14.1 to 19.4°C (Table 5).In a recent study at a pier overgrown by D. vexillum in the West Passage of Narragansett Bay, Rhode Island, recruits were first observed on settlement plates in July 2005 (Table 5); settlement occurred during a 7-day period of increasing water temperatures in the range of 19 to 20°C (GSO site of Auker 2006, Figures 20, 36;Auker and Oviatt 2008).Long term temperature records from the colonized areas of Georges Bank indicate that bottom temperatures do not exceed 17°C (Figure 6).Based on the foregoing data from coastal and offshore sites, we suggest that the release of brooded larvae and subsequent recruitment occur at water temperatures between 14 and 20°C and depend on local conditions.
At the Great Harbor (2007) and Damariscotta River ( 2006) sites where we have adequate temperature records (Figures 2, 4), highest recruitment of D. vexillum occurred during the  4 b GSO Dock site of Auker (2006, Fig. 36) warmest part of the year.This is also the case at the GSO site in Narragansett Bay (2005) mentioned above (Auker 2006, Fig. 36;Auker and Oviatt 2008).West of Narragansett Bay, at two sites in Connecticut (Mystic River and Avery Point), the highest recruitment occurred in the warmest part of the year in 2001 and 2003 but preceded the warmest season at Mystic River in 2002 (Osman and Whitlatch 2007, Figure 2).Recruits of D. vexillum continued to appear on the plates as waters cooled to below the temperature of first occurrence.At Great Harbor, the southern site, recruitment lasted for 5 months and ended in the range of 9.3 to 9.8°C.Colonies contained brooded larvae on January 4, 2008 (39 days after the last recruits appeared), indicating that cessation of recruitment does not necessarily mean colonies are empty of larvae.At Damariscotta River, recruitment lasted for at least 3.5 months and ended at or below 11.3°C but above 10°C.In Narragansett Bay, recruitment of D. vexillum lasted more than 3 months, from July 14 to October 21, 2005, but the experiment ended before the last recruits were observed (Auker 2006, Fig. 36;Auker and Oviatt 2008).At Mystic River and Avery Point in the years 2001 to 2003, recruitment lasted approximately 4 months (Osman and Whitlatch 2007).

Water temperature minima, maxima, and variability at the study sites
Water temperature minima in the cool season and temperature variability in the warm season are different at the study sites (Figure 1), and these factors seem to influence the recruitment of D. vexillum larvae.The Great Harbor site is the most southerly of the three study sites and displays the lowest temperatures in the cool season, followed by the Damariscotta River estuary site, and the Appledore Island site (Table 4).At Great Harbor, the coldest site, D. vexillum recruited at the highest temperature (18.7 to 19.4°C), followed by Damariscotta River (15.8 to 18.0°C), and Appledore Island (14.1 to 15.1°C), the mildest site (Tables 1-3, 5).At all three sites, temperature variability (within the well-defined seasonal trend) is low in the warm season (Figures 2, 4, 5).
The offshore gravel habitats of the Georges Bank study areas lie at 45 to 65 m water depth and experience milder water temperatures than the shallow coastal sites.In the areas colonized by D. vexillum (Ga, Gb), average daily bottom temperatures range from a low of 4°C in the cool season to a high of 17°C in the warm season (Figure 6).Thus, the minimum temperature on the bank is warmer than at Great Harbor and Damariscotta River, and for the most part Appledore Island (Table 4).As with the shallow water sites, temperature variability is low in the warm season at Ga and Gb (Figure 6).
Area Gc, the Georges Bank area free of D. vexillum, exhibits yearly minimum and maximum water temperatures similar to the colonized areas (Ga, Gb), but it displays high temperature variability (11°C) in the warm season (compare Figures 6 and 7).Gc is located on the bank edge where water temperature variability is high due to the movement and interaction of strong tidal currents, tidal fronts, and the mixing of warm unstratified bank water with highly stratified cool water masses of the bank edge and Gulf of Maine to the north (Loder et al. 1992).
The seasonal temperatures at Gc are well within the range tolerated by D. vexillum, and it is possible that since 2002, when the species was first observed in Ga, swimming larvae assisted by strong tidal currents could have colonized the 4 to 6 km of gravel habitat that separates Ga and Gb from Gc (Figure 1).In addition, Ga is trawled and dredged, and viable colony fragments could have been transported to Gc by currents and fishing activities.Although D.
vexillum can survive at Gc, and isolated colonies might occur, its failure to colonize the area likely is related to its lack of sexual reproduction due to the high variability of bottom temperatures in the warm season and the lack of a sustained warming trend.

Implications for the spread of Didemnum vexillum
Colonies of D. vexillum degenerate and all but disappear under extremely cold conditions in shallow water.We hypothesize that the degree to which colonies degrade in the cool season influences the length of time they require to regenerate, reproduce sexually, and brood and release larvae.Thus, larval recruits will be released at the end of a developmental period as water temperatures warm, not necessarily when a particular water temperature is reached.It follows that larvae likely will appear at different temperatures at climatically different sites, and at approximately the same time and temperature at climatically similar sites.The time required for a colony to develop and release larvae (not known at present) and the length of the warm season possibly affect the length of the recruiting period at a site.Larval recruitment ceased at Great Harbor in the range of 9.3 to 9.8°C, although mature non-feeding colonies still contained brooded larvae.Less definitive data from Damariscotta River indicate larval recruitment ceased in the range of 10.0 to 11.3°C.We speculate that as waters cool, larval release from healthy colonies is regulated chiefly by temperature and therefore should end at approximately the same temperature at all sites.Finally, at deeper water sites where minimum temperatures are warmer than at shallow sites, it is possible that D. vexillum colonies are not as affected in the cool season (i.e.do not degenerate) and thus could have a longer recruiting season.Although we have not observed the Georges Bank sites in the cool season, a long recruiting period could account for the highly successful spread of the species at those sites where the minimum temperature in the cool season is 4°C or higher and temperature variability in the warm season is low.
The results presented here and in cited studies can be applied to predict where D. vexillum will be able to reproduce sexually and develop colonies in marine habitats, given that substrate and food requirements of the species are met.
(1) Cold temperatures, at least to several degrees below zero centigrade, cause colonies to regress, but not all die.(2) The maximum temperature of survival of the species is not known but probably is higher than 25°C.(3) The first occurrence of recruits is not linked to a particular temperature, but likely depends on the length of time required by an overwintering colony to develop and release brooded larvae.We hypothesize that this time period depends on local temperature trends and on the condition of the colony at the end of the preceding cool season.The first occurrences of D. vexillum recruits at study sites in New England were observed at water temperatures in the range of 14 to 20°C.(4) Late in the warm season, larvae continue to recruit at temperatures below the temperature of initial appearance.Our data indicate larvae ceased to recruit in the range of 9 to 11°C.(5) During the warm season, highly variable temperatures likely inhibit the reproductive process and successful colonization.
Some types of sites are more susceptible to colonization than others.For example, estuaries and harbors with restricted circulation are settings where warm atmospheric temperatures can elevate temperatures of shallow waters, even at higher latitudes.These sites can experience higher temperatures in the warm season than headlands of nearby outer coasts where waters freely circulate and remain cooler.Island coasts surrounded by large bodies of water can experience milder winters than nearby mainland coasts.Deeper water sites can be more stable climatically than shallow water sites.Oceanographic processes such as tidal currents and movements of water masses can affect the variability of water temperatures at a site in the warm season.At present, D. vexillum occurs throughout New England and along the Maine coast to near the U.S./Canada boundary.The species has not yet been reported from Atlantic Canada.By examining seasonal temperature trends at coastal and offshore locations and applying the information presented here, it should be possible to identify sites at risk to be colonized by Didemnum vexillum.

Figure 1 .
Figure 1.Locations of study sites in New England waters of the Gulf of Maine region.Didemnum vexillum is present in coastal environments at Great Harbor (GH), Appledore Island (AI), and Damariscotta River (DR).The species is present offshore on Georges Bank gravel habitats (Ga, Gb) but not on a gravel habitat (Gc) 4-6 km north of the two colonized areas.Darkest line is the 100 m isobath.
, 4 c Darling Marine Center pier (DMC, 2008) d Buoy 5.1 km south of Appledore Island (UNH, 2008) e abnormally warm winter f partial record, approximate maximum temperature

Figure 2 .Figure 3 .
Figure 2. Average daily water temperature at 2 m (relative to MLLW) from January 1 to December 18, 2007 at Great Harbor.Temperature range of the record is -1.0 to 23.8°C.A, first occurrence of Didemnum vexillum recruits observed on June 26 (JD 177) at the end of a 4-day interval during which the temperature increased from 18.7 to 19.4°C.B, last occurrence of D. vexillum recruits observed on November 27 (JD 331) at the end of a 8-day interval during which the temperature decreased from 9.8 to 9.3°C

Figure 5 .Figure 6 .
Figure 5. Average daily water temperature at 1 m below water surface from January 23 to December 4, 2003 at a buoy located approximately 10 km off the New England Coast and 5.1 km south of Appledore Island, Maine (Figure 1).Temperature range of the record is 1.5 to 20.1°C.Graph is considered to be representative of seasonal temperatures at Appledore Island

Figure 7 .
Figure 7. Bottom temperature data compiled from 1977 to 2007 for an area (Gc) on Georges Bank free of Didemnum vexillum that lies 4-6 km north of two colonized areas (Ga, Gb).Temperature range of the record is approximately 4 to 16°C.Note the high variability of temperatures in the warm season (JD 210-300) and compare with Figure 6

Figure 8 .Figure 9 .Figure 10 .Figure 11 .
Figure 8. Didemnum vexillum oozooids and young colonies of the colonial ascidian Botrylloides violaceus (Bv).D. vexillum oozooids include the youngest recruit (1) with eye spots (es) and lateral organs of the thorax (LOTs) which are located in the anterior part of the individual and are not white as they have not yet produced spicules; several somewhat older oozooids (2) with small white LOTs (lot); and two older oozooids (3) with dense white LOTs (lot), a cloacal aperture (ca), oral siphon (os), and small white spicules distributed within the tunic.Great Harbor, MA, June 29, 2007, image 9237.Scale bar 2.0 mm

Table 2 .
Appledore Island, Maine.Observation dates, water temperatures, and occurrences of new Didemnum vexillum recruits.Four vertical settlement plates were deployed (d) at Appledore Island at a depth of 2.0 to 2.5 m.Plates were replaced with clean ones after each observation.Average daily water temperature recorded at plate depth

Table 3 .
Damariscotta River estuary, Maine.Observation dates, water temperatures, and occurrences of new

Table 4 .
Great Harbor, Massachusetts, Damariscotta River, Maine, and Appledore Island, Maine.Yearly minimum and maximum average daily temperatures from most recent older data for comparison with data from this study.No data (nd) indicates a partial temperature record a NOAA tide gauge station 8447930 (NOAA, 2008) b this study; see Figures 2, 3

Table 5 .
Summary of dates (Julian Days) and average daily water temperatures of first and last occurrences of Didemnum vexillum based on settlement plate studies in New England coastal waters.The interval of first or last occurrence columns show the number of days and the change in water temperature between observations of settlement plates