Observations of recruitment and colonization by tunicates and associated invertebrates using giant one-meter 2 recruitment plates at Woods Hole , Massachusetts

Large recruitment plates measuring 1 × 1 m were deployed over an 18-month period from September 2013 to March 2015 for the purpose of documenting recruitment and colonization processes of marine invertebrate species at Woods Hole, Massachusetts. Each side of two plates was subdivided into 16 subareas (25 × 25 cm), and an observational strategy was developed whereby, at approximately two-week intervals, a different subarea was cleaned. Using this approach, we were able to photographically document species recruitment and growth interactions. Water temperature records from the site show that steady warming and cooling between 3 and 20° C changed at a mean rate of 0.2 ° C d. However, temperature changes during the coolest and warmest parts of the temperature cycle were highly variable. In 2014, between the first and last occurrence of 0° C, temperatures were ≤0° C 15 percent of the time, but in 2015 temperatures were ≤0° C 93 percent of the time. In 2014, between the first and last occurrence of 21° C, temperatures were ≥21° C 88 percent of the time, and this warm period correlated with the disappearance of the hydroid Ectopleura crocea, the solitary tunicates Ascidiella aspersa and Ciona intestinalis, and the 2013 generation of Botrylloides violaceus. In Woods Hole, large plates provided enough space to accommodate both fastand slow-colonizing species, resulting in the establishment of a diverse assemblage that was observed over a long time period. The most successful colonizing species had relatively long reproductive and recruitment periods, grew rapidly, repelled settlement onto their surfaces by larvae of any species, defended themselves against overgrowth by any species, overwintered, and lived a long time. Of the three dominant species observed in this study, the colonial tunicates Didemnum vexillum and Botrylloides violaceus had these qualities; the encrusting colonial bryozoan Schizoporella unicornis had all but one, it grew more slowly than the others. Barnacles constituted the only biological substrate that was effectively colonized by other species, both by larval recruitment and overgrowth. In Woods Hole, after a substrate had become fully colonized, there was very little opportunity for new recruitment or colony growth until new substrate opened after the death of colonies and individuals and the disappearance of biogenic structures such as amphipod tubes. An understanding of colonization processes utilized by invasive species allows prediction of their potential effects on ecosystems in areas where they are not yet present.


Introduction
Previous experiments using small-size substrates at the study site in Woods Hole were directed toward studying the invasive colonial ascidian Didemnum vexillum Kott, 2002 to determine the timing of larval recruitment and development, colony response to low water temperatures in winter, and the ability of colony fragments to re-attach to hard substrates after weeks of suspension in the water column.The experiments used small (12 × 12 cm, 144 cm 2 ), horizontallyoriented, recruitment plates (Valentine et al. 2009) or small plastic containers (base = 38.5 cm 2 ) (Morris and Carman 2012) deployed from a dock.The small recruitment plates were large enough for collecting information on the behavior of small colonies of D. vexillum.However, because of the very aggressive growth and dominance of this species, they were too small to document the occurrences and interactions of associated species.Osman (1977) reached a similar conclusion on the inadequacy of small recruitment plates in this regard during a long-term recruitment study near Woods Hole.We decided to use 1 × 1 m recruitment plates in the expectation that a large sampling area would allow us to document the recruitment, longevity, and inter-and intraspecies interactions of a range of species, including D. vexillum, that are common constituents of the colonizing assemblage at the study site.We developed a methodology for observing and photographically sampling subareas of the plates that we pursued over two winters and one summer from early September 2013 to late March 2015.Our results are based on qualitative field observations that revealed patterns of species behavior.There is great potential to quantify species interactions using large plates.

Study site
The study site is located in an open area of the dock at the Woods Hole Oceanographic Institution (41.524°N, 70.672° W) in Great Harbor, Woods Hole, Massachusetts (Figure 1).The opening (~3.6 m on a side) is used for testing equipment and is serviced by a stationary crane.Water depth is approximately 16 m.The study site is bounded on 4 sides by pilings of cement and steel which are colonized by attached epifauna, including mussels, barnacles, bryozoans, tunicates, sponges, hydroids, and anemones, among others.The National Oceanic and Atmospheric Administration (NOAA) maintains a tide gauge station (8447930) and water temperature logger (1.8 m below mean low water) at the dock (NOAA 2016).The mean tidal range is 0.5 m, and maximum tidal currents range up to 1 m s -1 .The temperature record used here (hourly mean of observations at six-minute intervals) extended from September 1, 2013 through June 1, 2015, a period of 639 days.The water column at the site is fully mixed to at least 10 m water depth (Valentine et al. 2009).

Configuration and handling of the one-meter 2 recruitment plates
Recruitment plates were 1 × 1 m sheets of medium gray PVC with a thickness of 6.35 mm (Figures 2, 3).Two plates were used in this study.The size of the large plates required they be deployed vertically for ease of deployment and retrieval and to minimize resistance to tidal currents at the study site.A hole was drilled into the upper left and lower left corners of the plate.A stainless steel shackle with a swivel was fitted into each hole.A rope from the upper left swivel extended up through the water to the dock, and a recruitment plate into sixteen subareas, each measuring 25 × 25 cm (625 cm 2 ).During a hypothetical experiment, eight white areas are cleaned sequentially through time, while eight shaded areas remain uncleaned.Subareas are distributed so that each cleaned subarea is bounded by one cleaned and two uncleaned subareas.See the text for an explanation of the observational strategy used in this study.
rope from the lower left swivel extended downward to a heavy weight (14 kg).When deployed, the vertical plate aligned with the tidal current direction.Plate movement in the water was analogous to a flag on a pole responding to wind currents.At retrieval, the plate was raised out of the water by the lifting rope using an overhead crane.The weight was unhooked, and the plate was lowered into a horizontal position and submerged in a plastic tray 5.5 cm deep holding circulating seawater provided by a submersible pump that hung from the dock.The plate was exposed to air for 30 to 45 s during transfer from the water to the tray.Four stainless steel bolts protruded 1.5 cm from near the corners of each side of the plate and acted as legs to prevent the crushing of invertebrates attached to the downfacing side of the plate when it lay horizontally in the tray.The water in the tray covered the fauna attached to the upper surface of the plate.Upon placement in the water-filled tray, a 1 × 1 m wooden frame (subdivided into 16, 25 × 25 cm subareas by wires stretched across the frame) was placed around the plate.Photographs were made of each 25 × 25 cm subarea on both sides of the plate, and of species interactions and recruits.The entire process required approximately one hour to complete for each plate.There is an overhead crane at the study site, but as the plate when fully covered by attached organisms weighed only ~40 kg without the weight, a smaller lifting device would suffice.Lifting and manipulating the large plate by hand would be difficult.

Sampling strategy and analysis
The four phases of colonization by sessile invertebrates described by Keogh and Downes (1982) are a) larval development and dispersal, b) testing by larvae for suitable substrate on which to settle, c) larval settlement (attachment and metamorphosis), and d) survival as a larva or juvenile until identified by an observer, at which time organisms become recruits.We view the colonization of substrates to be a step-wise process that begins with the four phases described above followed by e) the occupation of a substrate area through growth, f) growth and larval-settlement interactions with other living species, and g) a reproductive phase.In addition, some species "colonize" a substrate by building structures, e.g.tubes.Here, recruits are identifiable juvenile organisms and small colonies that have settled and survived during the period between observations.Growth inhibitions called standoffs develop when neither of two colonial species can overgrow the other.Species in a standoff commonly develop ridges of tissue at the interface between the neighboring colonies.Another kind of standoff occurs when the growth of a colonial species is inhibited by movements of other organisms; for example the moving appendages of barnacles, bryozoans, and hydroids, the movements of gastropods, and the flow of water by the siphonal currents of solitary tunicates.Some colonial species compete with other species by overgrowing and killing them or by partially overgrowing them and inhibiting their growth.The term overwintering refers to an individual or colony that ceases growth when water temperatures cool and survives to resume growth when temperatures warm.A regressing colony shows signs of tissue deterioration that can occur at any time of the year, depending on the life-history of the species.The large plates afforded the opportunity to observe settlement and recruitment of species, seasonal occurrences and longevity of species, and species interactions through time over a single surface.In order to record the timing of recruitment, the plate surfaces needed to be photographed and cleaned periodically.We used two plates, A and B, and the two sides of each plate were labeled Front and Back for purposes of identification.We used photographic images to qualitatively assess colonization processes.It B. Plate AF removed from water on June 30, 2014, 299 days after initial submergence, 66 days after cleaning 8 subareas (Figure 2A), and 14 days after subarea 9 was cleaned for the second time (Table 1 is beyond the scope of this initial study using large recruitment plates to provide a quantitative analysis of species recruitment, growth rates, and competition. On September 4, 2013, in preparation for observations of recruitment in early 2014, both sides of the two plates (4 m 2 ) were thoroughly cleaned (scraped and scrubbed) and suspended from the dock at a depth of 1 to 2 m below mean low water (Table 1, observation 0) for the purpose of establishing an assemblage of attached species that would recruit in late fall and overwinter on the plates.By April 2014, both sides of the plates were colonized by attached species.The fauna on one side of each plate (A Back, B Front) was observed until November 5, 2014.The other side of each plate (A Front, B Back) was divided by the removable grid into sixteen, 25 × 25 cm subareas for photographic documentation of colonization processes (Figure 2).Eight subareas (2,3,5,8,9,12,14,15) on A Front and B Back were selected to record recruitment and growth interactions during the experiment.They were cleaned on April 25, 2014 (Table 1, observation 5).Each of the 8 cleaned subareas was bounded by one cleaned subarea and two uncleaned subareas that were colonized beginning in September 2013.The other 8 subareas (1,4,6,7,10,11,13,16) on A Front and B Back were not cleaned.Potential sources of larvae for recruitment beginning in April 2014 included the heavily colonized dock pilings on all sides of the study site, the 8 colonized subareas on one side of each plate (AF, BB) that were not cleaned, and the colonized reverse side of each plate (AB, BF).At subsequent observations of the plates, beginning with plate AF, one subarea that had been cleaned on April 25, 2014 (observation 5) was re-cleaned (except for subarea 2 that was cleaned on April 25 but not again).For example, subarea 3 was cleaned at observation 5 and recleaned on the next visit (observation 6) but not  1, 3).Hourly data is from National Oceanic and Atmospheric station 8447930 located at the study site (NOAA 2016).The record has been smoothed by plotting every 10 th hourly data point.Triangle symbols represent the first day of a month.again (Table 1).After a subarea had been cleaned, it was checked for recruitment on the following visit.As a further example, subarea 9 on plate AF was cleaned at observation 5 and checked for recruitment at observation 6; it was next cleaned at observation 9 and checked for recruitment at observation 10.This strategy allowed recruitment to be observed twice in subarea 9 and colonization to be documented from observation 6 to 9 and from observation 10 to the end of the experiment.A total of 14 subareas (7 on plate AF followed by 7 on plate BB) were cleaned from May 5 to November 5, 2014.On each visit, observations were photographically recorded in specific subareas, as described above, and in other parts of the plates as needed.
After the initial deployment of the plates on September 4, 2013, a total of 22 observations of the recruitment plates were made during the experiment.The plates were inspected twice in September and once in October of 2013, and again on April 14, 2014 to determine the status of overwintering species (Table 1).Fifteen observations were made at approximately 14-day intervals from mid-April to early November 2014.In the description of results that follows, the reader should note that the dates and water temperatures associated with recruitment and colony growth events are those recorded at the end of an observation period.For example, the first observed recruitment of D. vexillum in 2014 was on June 30 when the water temperature was 21.0° C. The first recruit actually settled sometime between June 16 and June 30 when temperatures increased from 18.3 to 21.0° C. It is not possible to more accurately determine the timing of such events without making observations at shorter time intervals.

Water temperature
The annual water temperature cycle from late 2013 to mid-2015 at Woods Hole plotted as an approximately symmetrical curve with small superimposed temperature fluctuations caused by local atmospheric variability (Figure 4).A pattern of slow, seasonal rise and fall of hourly water temperatures characterized most of the year.Water temperatures between 3 and 20° C increased and decreased at a steady pace.Analysis of the one warming period (~ April to ~July, 88 d) and the two cooling periods (~Oct to ~Jan, 108 d each) determined that temperatures increased at a mean rate of 0.19° C d -1 and decreased at an almost identical rate of 0.16° C d -1 .
Water temperatures were more variable during the coolest and warmest parts of the year (Table 2).During the coolest part of the year in 2014, Table 2. Timing and variability of water temperature during coolest and warmest periods of the annual temperature cycles in 2014 and 2015.Time periods are determined by the first and last observations of temperature values that define a range.For example, during the development of the cool period in 2014, hourly temperature values ≤ 0.0° C (the defined range) occurred from January 24 to March 6, a total of 42 days, during which 983 out of a possible 1008 hourly observations were recorded.During the time period, water temperature varied considerably, as only 148 (15 percent) of the temperature observations were ≤ 0.0° C (Figure 4).Both maximum and mean temperatures were >0.0° C. Hourly temperature data is from National Oceanic and Atmospheric Administration station 8447030 located at the study site (NOAA 2016).The number of hourly temperature observations is up to 3 percent less than expected due to gaps in the data.

Occurrence and recruitment of major colonizing species
The following species were major constituents of the assemblage that developed on the plates in Woods Hole from September 2013 to March 2015.The presence of settled, recently-metamorphosed recruits was used to determine the recruitment dates and temperatures for the colonial tunicates D. vexillum and Botrylloides violaceus Oka, 1927, and the colonial bryozoan Schizoporella unicornis (Johnston, 1874).Barnacle species recruited to the plates throughout the study period.They played a role in colonization of the plates, but they could not be identified to species for this experiment.Early post-settlement recruits were not observed for the colonial hydroid Ectopleura crocea (Agassiz, 1862), so its recruitment dates were based on the first occurrence of small colonies.The occurrence of the amphipod Jassa marmorata Holmes, 1905 was based on the presence of individuals and their tubes.In the following descriptions, observation dates are followed by temperature values that represent the mean water temperature recorded on that day.
In 2014, recruitment of D. vexillum colonies was observed on June 30 (21.0° C) and continued for sixteen weeks to October 20 (16.8° C).Peak recruitment was observed on July 14 (22.1°C) and July 28 (22.3°C).Colonies were not inspected for larval content.Potential sources of early recruits were D. vexillum colonies from 2013 that had overwintered on the recruitment plates and on the adjacent dock pilings.Larvae that settled later in the year could have been produced by the overwintering colonies just mentioned and by the colonies that recruited early in 2014 to the plates and dock pilings.Thus, by June 30, 2014 two generations of D. vexillum colonies occupied the plates, those that recruited in 2013 and those that recruited in 2014 (Table 3).Some of the colonies that overwintered from 2013 and began growing by May 5, 2014 began to regress by August 25 (20.9°C) while others continued to grow until at least January 21, 2015 (1.8° C).The tunics of the regressing colonies were pale white and exhibited dark areas that represented cloacal canals clogged by fecal pellets (Valentine et al. 2007).Colonies that recruited in 2014 had a shorter growing season than those that recruited in 2013 and were smaller.They showed the first signs of regression by November 5 (12.6°C),   (Milne-Edwards, 1841).Growth interactions between D. vexillum and S. unicornis resulted in standoffs, except that D. vexillum colonies partially overgrew the margins of regressive S. unicornis colonies.The species also partially overgrew the tunics of solitary tunicates, including Ciona intestinalis (Linnaeus, 1767), Styela clava Herdman, 1881, and Ascidiella aspersa (Müller, 1776) but was unsuccessful in closing off their active siphons.Growth at the margins of D. vexillum colonies sometimes was prevented by the movements on the plates of the filter-feeding gastropod Crepidula fornicata (Linnaeus, 1758).

Botrylloides violaceus (invasive colonial tunicate)
After recruitment plates were cleaned on September 4, 2013 (21.7°C), newly-recruited B. violaceus colonies were observed from September 20 to October 18 (19.7 to 17.2° C), the last observation in 2013 (Table 3).At the next observation of the plates on April 14, 2014 (6.9° C), small B. violaceus colonies were present that had recruited in late 2013.They had increased in size since they were last observed in 2013, and apparently they were growing on April 14, as they had increased in size by April 25 (8.9°C), 11 days later.
In 2014, recruitment of B. violaceus colonies was observed on June 16 (18.3°C) and continued for twenty-four weeks to December 1, 2014 (8.2° C).Peak recruitment was observed on June 30 (21.0° C) and July 14 (22.1°C).Colonies were not inspected for larval content.Potential sources of the early-settled larvae were B. violaceus colonies from 2013 that had recruited to plates and to the adjacent dock pilings.Some colonies that overwintered from 2013 began to regress (fall apart) by June 30, 2014 (21.0°C) during peak larval recruitment, and all of the 2013 colonies had disappeared by August 25 (20.9°C).Thus, larvae that settled later in 2014 were produced by colonies that recruited to the plates and dock pilings earlier in 2014.The colonies continued to grow until at least January 21, 2015 (1.8° C), but were not growing and showed no signs of regression when observed on March 26, 2015 (1.1° C).
Larvae of B. violaceus successfully recruited to the plate substrate and to the tests of barnacles (Table 4).Larvae did not recruit to the living tissues of itself and other colonial tunicates, solitary tunicates, and the encrusting colonial bryozoan S. unicornis.Growth interactions between recently-settled recruits of B. violaceus and D. vexillum resulted in standoffs.Colonies of B. violaceus successfully overgrew unoccupied plate substrate, recently-settled recruits of S. unicornis and most, but not all barnacle tests, amphipod tubes, and hydroid stalks (except for some polyps at the ends of stalks).Growth interactions with colonies of itself, other colonial tunicates, and S. unicornis resulted in standoffs, except that B. violaceus colonies partially overgrew the margins of regressive S. unicornis colonies.The species also partially overgrew the tunics of solitary tunicates but was unsuccessful in closing off their active siphons.

Schizoporella unicornis (invasive colonial bryozoan)
After recruitment plates were cleaned on September 4, 2013 (21.7°C), newly-recruited S. unicornis colonies were observed from September 20 to October 18 (19.7 to 17.2° C), the last observation in 2013 (Table 3).At the next observation of the plates on April 14, 2014 (6.9° C), small S. unicornis colonies were present that had recruited in late 2013 and overwintered.They had increased in size since they were last observed in 2013.However, they were not growing, as they had not increased in size by April 25, 11 days later.The non-growing colonies were a pale white color.They began to grow by May 28 (14.8°C) and gradually changed color from white to pink to pale orange.
In 2014, recruitment of S. unicornis colonies was observed on June 16 (18.3°C) and continued for 40 weeks to at least March 26, 2015 (1.1° C).A period of peak recruitment was not observed.
Colonies were not inspected for larval content.Potential sources of the early-settled larvae were S. unicornis colonies from 2013 that had overwintered on the recruitment plates and on the adjacent dock pilings.Larvae that settled later in 2014 could have been produced by the overwintering colonies just mentioned and by colonies that recruited earlier in 2014 to the plates and dock pilings.Thus, by June 16, 2014 two generations of S. unicornis colonies occupied the plates, those that recruited in 2013 and those that recruited in 2014 (Table 3).Some of the overwintered colonies from 2013 began to regress by September 8 (22.5°C) while others continued to grow until at least October 20 (16.8° C).All growth had ceased by January 21, 2015 (1.8° C).Colonies that recruited in 2014 had a shorter growing season than those that recruited in 2013, and they did not show the first signs of regression until November 5 (12.6°C), later than the 2013 colonies.
Larvae of S. unicornis successfully recruited to the plate substrate and to the tests of barnacles (Table 4).Larvae did not recruit to the living tissues of itself or of solitary and colonial tunicates.Colonies successfully overgrew unoccupied plate substrate and most, but not all barnacle tests.Growth interactions with colonies of itself, D. vexillum and B. violaceus resulted in standoffs.Some barnacles recruited to the surface of S. unicornis when it was overwintering.

Ectopleura crocea (native colonial hydroid)
Here we refer to this species as E. crocea, but as Schuchert (2010) notes, it is very difficult to distinguish from E. larynx (Ellis and Solander, 1786).After recruitment plates were cleaned on September 4, 2013 (21.7°C), E. crocea recruits were first observed on May 28, 2014 (14.8°C) (Table 3).Recruits and large growing colonies were observed thereafter until June 30 (21.0° C).Colonies began to regress by July 14 (22.1°C), and they disappeared by July 28 (22.3°C).Potential sources of recruits were E. crocea colonies that had overwintered from 2013 on adjacent dock pilings.Growing colonies reappeared on the plates by October 20 (16.8° C) and were present until March 26, 2015 (1.1° C).Although colonies of E. crocea rapidly colonized open areas of the plates, they were not observed to recruit to or overgrow any other species.By contrast, they were partially and sometimes completely overgrown by colonies of D. vexillum and B. violaceus (Table 4).

Jassa marmorata (cryptogenic amphipod)
After recruitment plates were cleaned on September 4, 2013 (21.7°C), J. marmorata individuals, but no tubes, were observed on October 18 (17.2°C), at which time barnacles of various sizes covered much of the plate area (Table 3).At the next observation on April 14, 2014 (6.9° C), the plates were covered by the brown J. marmorata tubes attached to the plate substrate and to the barnacle substrate that was developing in October, 2013.Individual amphipods were observed in and on the tubes, which did not increase in number over a period of at least 2.5 months until June 30, 2014, after which the tubes fell off the plates.Both amphipods and their tubes were observed in small numbers on January 21, 2015 (1.8° C) and March 26 (1.1° C).
The amphipod tubes that were abundant when first observed in April 2014 were not attached to the colonies of D. vexillum and B. violaceus that recruited in 2013 (Table 4).However, their tubes partially to fully covered barnacles and parts of the tunics of the solitary tunicates A. aspersa and C. intestinalis.Colonial tunicates D. vexillum, B. violaceus, and A. glabrum overgrew J. marmorata tubes to a large extent except in a few instances where tubes occupied by amphipods remained open.Barnacle larvae did not settle and recruit to the tubes.

Barnacle species
Barnacles produce planktonic larvae that are present in the water column in Woods Hole throughout most of the year (Fish 1925;Pineda et al 2002).After recruitment plates were cleaned on September 4, 2013 (21.7°C), newly-recruited barnacles were observed on September 20 (19.7° C) and on October 18 (17.2°C) (Table 3).Observations in 2014 began on April 14 (6.9°C), and barnacle recruits were first observed on May 5 (9.7°C) and then on May 14 (12.0°C).No further recruitment was observed until July 14 (22.1°C), after which recruits were consistently recorded over an 8-month period until March 26, 2015 (1.1° C).At times many recruits appeared within a 2-week period and at other times they were few in number, possibly reflecting the reproductive cycles of several different species.Potential sources of barnacle recruits that settled in late 2013 and early 2014 were barnacles on the adjacent dock pilings.Larvae that settled later in the year could have been produced by barnacles on the plates and on the dock pilings.
Barnacles recruited almost exclusively to clean plate surfaces and were not observed on colonial tunicates and amphipod tubes, or on most solitary tunicates except for A. aspersa, C. intestinalis, and S. clava (Table 4).They did not recruit to the colonial bryozoan S. unicornis except when it was overwintering.Growth interactions between closely-spaced barnacles resulted in deformation of their test margins.Larvae of D. vexillum and B. violaceus recruited to the sides of barnacle tests.Most barnacles were overgrown by colonies of D. vexillum, B. violaceus, D. listerianum, and S. unicornis, and the tubes of J. marmorata.

Occurrence of minor colonizing species
The following species were minor constituents of the assemblage that recruited to the plates in Woods Hole from September 2013 to March 2015.Settled, recently-metamorphosed recruits were not observed, probably because of their scarcity, so recruitment dates and temperatures are based on the presence of partially-grown colonies and individuals.In the following descriptions, observation dates are followed by temperature values that represent the mean water temperature recorded on that day.

Aplidium constellatum (native colonial tunicate)
Colonies of A. constellatum were observed during a 3-month period from August 11 to October 20, 2014 (22.0 to 16.8° C) (Table 5).Regressing colonies were observed on November 5 and December 1 (12.6 to 8.2° C).Recently-settled recruits were not observed during the experiment.Growth interactions with itself, D. vexillum, and A. glabrum resulted in standoffs (Table 4).No larvae of D. vexillum, B. violaceus, S. unicornis, or barnacles were observed to recruit to A. constellatum.

Aplidium glabrum (native colonial tunicate)
Colonies of A. glabrum were observed during a 9-month period from June 16, 2014 to March 26, 2015 (18.3 to 1.1° C) (Table 5).Recently-settled recruits were not observed during the experiment.Growth interactions with itself and the other colonial tunicates resulted in standoffs (Table 4).The species also displayed standoffs with growing S. unicornis, but partially overgrew the margins of overwintering colonies of that species.Aplidium glabrum partially to fully overgrew the tubes of the amphipod J. marmorata.No larvae  2009, and references therein).Skeleton shrimp (Table 5) did not colonize the plates in terms of occupying space, but they were present in large numbers during a 3-month period from April 14 to July 14, 2014 (6.9 to 22.1° C).After an approximate 3-month hiatus, they were observed from October 20, 2014to March 26, 2015 (16.8 to 1.1° C), the end of the experiment.

Discussion
Observations of two 1 × 1 m recruitment plates over a period of 18 months and a sampling strategy that required the periodic cleaning of 25 × 25 cm subareas of the plates showed that successful colonization by several species likely was controlled by a combination of factors, including: a) available firm substrate; b) water temperature; c) timing and duration of recruitment; d) timing and duration of growth; e) the ability of organisms to overgrow other organisms and/or to defend against overgrowth; and f) the ability of species to maintain their space by overwintering and living longer than competitors.The interplay of these factors resulted in the dominance, over varying periods of time, of two colonial tunicates (D. vexillum, B. violaceus), a colonial bryozoan (S. unicornis), a colonial hydroid (E.crocea), a tube-building amphipod (J.marmorata), and barnacle species.The recruitment plates were isolated from the seabed and were not affected by sediment movement or by the activities of species living on the seabed.Tidal currents did not exceed 1 m s -1 .Substrate and water temperature provided the physical framework within which recruitment and colonization occurred.The principal substrate in this study was the firm surface of the PVC plates and the tests of living and non-living barnacles.Plate surfaces became available for colonization through periodic cleaning of selected subareas of the plates, the death of colonial species (B.violaceus, E. crocea), and the falling off of amphipod tubes and dead barnacle tests.
The initial colonizers of the plates, who overwintered from 2013 to 2014, were chiefly barnacles and tube-building amphipods, and to a minor extent three colonial species, two tunicates (D. vexillum, B. violaceus) and one bryozoan (S. unicornis).None of these species recruited to the plates from April 25 to May 28, 2014, during which time water temperature increased from 8.9 to 14.8° C.This recruitment gap possibly began earlier in the year, but we have no observational data for that time period.
If we define the cool season as occurring between the first and last occurrences of 0° C, the differences between 2014 and 2015 in duration (days), temperature variability, and temperature minima did not seem to affect colonization by the dominant colonial tunicates D. vexillum, and B. violaceus, the colonial bryozoan S. unicornis, and the hydroid E. crocea, all of which successfully overwintered in 2015 when temperatures remained ≤-1° C for several weeks.Sustained mean temperatures above 21° C in 2014 possibly contributed to the disappearance of the colonial hydroid E. crocea, the tube-building amphipod J. marmorata, caprellid amphipods, and the less well-documented solitary tunicates, C. intestinalis and A. aspersa.Temperature and/or senescence may have contributed to the disappearance of the 2013 generation of the colonial tunicate B. violaceus.The disappearance of the tunicates, the hydroid, and the amphipod tubes made substrate available to recruits of other species.
Didemnum vexillum was a very successful colonizer.Its strongest qualities included rapid colony growth, overgrowth of some competitors for space (E.crocea, J. marmorata tubes, barnacle tests), and the ability to defend itself by preventing larvae from settling onto its surface grown.The amphipods and their tubes disappeared by late June (~21° C) as water temperatures warmed, offering substrate for recruitment by other species.
Barnacle species recruited to unoccupied substrate during most of the year, sometimes in high numbers, but they were not effective in maintaining their presence over a long period.They did not recruit to the surfaces of any other species except occasionally to the tunics of A. aspersa, C. intestinalis, and S. clava, and to the surfaces of overwintering S. unicornis colonies, and they were ineffectual in overgrowing these species.Barnacles were overgrown by D. vexillum, B. violaceus, and S. unicornis and covered by the tubes of the amphipod J. marmorata.Barnacles produced many larvae in most seasons and, in the absence of competitors for space, had the potential to be successful colonizers.
The minor colonial tunicate species were represented by few individuals, but they showed consistent and interpretable patterns of presence and absence.Larval recruits were not observed for these species, but our observations suggest they require unoccupied substrate for settlement.Colonial tunicates A. constellatum, A. glabrum, and B. schlosseri were first observed in the warm season (mid-June to mid-August) of 2014 and persisted at least into November, with A. glabrum overwintering into late March 2015.These species, though minor in terms of area colonized, were successful in maintaining their presence by creating standoffs with other colonial tunicates, including the faster-growing D. vexillum and B. violaceus, and by not allowing settlement onto their surfaces by larvae of other species.In one exception, D. vexillum was observed to overgrow a small colony of B. schlosseri.In contrast to its minor role in the present study, A. glabrum was shown to be a major colonizer of a complex subtidal vertical rock wall substrate in Nahant, MA (Sebens 1986), although potential competitors D. vexillum and B. violaceus were not reported to be present at the site.Another colonial tunicate, D. listerianum, covered much of the plate surfaces in September and October 2013 at the start of the experiment but was not observed thereafter.
Of the solitary tunicates recorded during the experiment, A. aspersa, C. intestinalis, and Styela clava first appeared in April 2014, possibly the result of recruitment in the latter part of 2013 or the early part of 2014.Two of these species, A. aspersa and C. intestinalis likely are sensitive to warm water temperatures in Woods Hole.Aplidium aspersa was not present after mid-July but re-appeared in March 2015.Ciona intestinalis was not present after mid-August 2014 but reappeared in early October and persisted to late March 2015.There is some evidence that northern populations of C. intestinalis in Denmark do not tolerate warm temperatures well, possibly because filtration rates decrease markedly above 21° C (Petersen and Riisgard 1992).Styela clava apparently is less sensitive to temperature change, as it was present from mid-April 2014 to late March 2015.Solitary tunicates served as substrate for overgrowth by D. vexillum and B. violaceus, although overgrowth was incomplete and did not impede the flow of water through their inhalant and exhalent siphons.
In Woods Hole, once a substrate became fully colonized, there was very little opportunity for new recruitment until new substrate opened after the death of colonies and individuals and the disappearance of structures such as amphipod tubes.The most successful colonizer of the recruitment plates in Woods Hole was D. vexillum followed by B. violaceus.Both species had relatively long reproductive and recruitment periods, grew rapidly, repelled settlement by larvae of any species, defended themselves against overgrowth by any species, overwintered, and lived a long time.Peak recruitment for both species occurred early in their reproductive seasons.Botrylloides violaceus reproduced over a longer time period and overwintered without regressing, but its colonies were shorter-lived than those of D. vexillum.We do not yet know if some portion of B. violaceus colonies always dies in the warm season in Woods Hole.Didemnum vexillum colonies grew more rapidly, and they overgrew new recruits of B. violaceus and S. unicornis.The bryozoan S. unicornis was a long-lived species but was a somewhat less successful colonizer because it grew more slowly than D. vexillum and B. violaceus.However, it had a longer recruitment period, which provided more opportunity for it to recruit to newly-available substrates.In addition, the calcareous, encrusting colonies of S. unicornis were physically harder and more firmly attached to the substrate than those of D. vexillum and B. violaceus.When comparing these three species, it appears that D. vexillum and B. violaceus could be better adapted to rapid, long-term colonization and S. unicornis to slow, long-term colonization.
Among the remaining species observed, the colonial hydroid E. crocea was very successful in colonizing open plate substrate, but was not able to recruit to substrate occupied by other species, and its longevity was shortened apparently by warming water temperature.Ectopleura crocea colonies and J. marmorata tubes disappeared from the plates and opened recruitment space for D. vexillum and B. violaceus during their peak recruitment periods.The colonial tunicates A. glabrum and B. schlosseri, when compared to D. vexillum and B. violaceus, grew relatively slowly, and their size was constrained by standoffs with faster-growing neighbor species.Solitary tunicates provided substrate for colonial tunicates, as their tunics did not repel attachment, but their siphonal currents did prevent complete overgrowth.Barnacle tests constituted the only biological substrate that was successfully colonized by other species.
The size of recruitment plates determines their effectiveness in providing a record of substrate colonization by attached invertebrates.Small plates (e.g.10 × 10 cm) are suitable for documenting settlement and recruitment of larvae over short time periods but often are too small to record processes of colonization, especially in the presence of rapidly-growing invasive species such as D. vexillum and B. violaceus.Large plates (1 x 1 m) permit the design of innovative strategies for sampling and cleaning areas of the substrate so as to observe the timing and environmental conditions of species recruitment, as well as recruitment and growth interactions during the colonization process.In Woods Hole, large plates provided enough space to accommodate both fast-and slow-colonizing species, resulting in the establishment of a diverse assemblage that was observed over a long time period.Large plates potentially are suitable for monitoring the effects of climate on the growth of species and on the stability of assemblages over time.The methodology described here can be used to document the longevity of colonial species, which to date has been difficult to determine.Finally, an understanding of colonization processes utilized by invasive species allows prediction of their potential invasive success and subsequent effect on ecosystems in areas where they are not yet present.

Figure 1 .
Figure 1.Maps showing the location of the study site at the dock of the Woods Hole Oceanographic Institution in Woods Hole, Massachusetts.

Figure 2 .
Figure 2.Diagram showing the division of the surface of a 1 m 2 recruitment plate into sixteen subareas, each measuring 25 × 25 cm (625 cm 2 ).During a hypothetical experiment, eight white areas are cleaned sequentially through time, while eight shaded areas remain uncleaned.Subareas are distributed so that each cleaned subarea is bounded by one cleaned and two uncleaned subareas.See the text for an explanation of the observational strategy used in this study.

Figure 3 .
Figure 3. Views of 1 × 1 m PVC recruitment plate AF (plate A, front side).Plate is submerged in holding tray; white hose (upper right) delivers seawater.A grid of orange wires lies on the plate and divides it into sixteen, 25 × 25 cm subareas for photography.See Figure 2 for layout of subareas.Two plates (A and B) were deployed on September 4, 2013 to begin the experiment.Compare subareas of the plate over time to see colonization (photographs by D. Blackwood). A. Plate AF removed from water on April 25, 2014 after 233 days of submergence, showing eight subareas cleaned on this date.Colonized areas are covered by brown tubes of the amphipod Jassa marmorata.Red patches are growing colonies of Botrylloides violaceus that overwintered from 2013.White patches in center are overwintering, non-growing colonies of Didemnum vexillum.B.Plate AF removed from water on June 30, 2014, 299 days after initial submergence, 66 days after cleaning 8 subareas (Figure2A), and 14 days after subarea 9 was cleaned for the second time (Table1).Image shows growth of red B. violaceus and pale orange D. vexillum colonies that overwintered from 2013.First recruits of D. vexillum were recorded on this date.Dark areas are Ectopleura crocea colonies covering parts of the eight cleaned subareas shown in Fig. 2A.Note growth of D. vexillum and B. violaceus into cleaned subareas.C.Plate AF removed from water on July 28, 2014, 327 days after initial submergence, 94 days after cleaning 8 subareas (Fig.2A), and 14 days after subarea 14 was cleaned for the second time.Image shows rapid growth of pale yellow D. vexillum and regression of red B. violaceus colonies that recruited in 2013.Dark areas are dead polyps of E. crocea.

Figure 4 .
Figure 4. Water temperature record from September 1, 2013 to June 1, 2015 at the study site located at the dock of the Woods Hole Oceanographic Institution in Woods Hole, MA.Numbered boxes indicate observations when large recruitment plates were retrieved from the water and photographed (see Tables1, 3).Hourly data is from National Oceanic and Atmospheric station 8447930 located at the study site (NOAA 2016).The record has been smoothed by plotting every 10 th hourly data point.Triangle symbols represent the first day of a month.

Table 1 .
Strategy for documenting colonization processes using large recruitment plates at Woods Hole, Massachusetts.Two verticallyoriented 1 × 1 m recruitment plates (A, B) were cleaned and deployed on September 4, 2013, observation 0. Subarea cleaning and observation at approximately two-week intervals began at observation 5.The sampling strategy for one side of each plate (A Front, B Back) is shown below.See the text for a full explanation of the sampling strategy.Figure2shows the arrangement of cleaned and uncleaned subareas on plates AF and BB.Cleaning of individual subareas ended on November 5, 2014 with subarea 14 on plate BB.Twenty-two observations continued to March 26, 2015 and are plotted on the temperature curve in Figure4.Abbreviations: c, subarea cleaned (shaded); Days, time between observations; Obs., observation number (plates removed from water and photographed).

Table 3 .
Recruitment and growth of major colonizing species.Record of species' colonization of two 1 × 1 m recruitment plates from September 4, 2013 to March 26, 2015.Observations 1 to 22 are plotted on the temperature curve of Figure 4.Note gap in recruitment of D.vexillum, B. violaceus, and S. unicornis in early 2014.See text for explanation of individual species' recruitment and growth patterns and Table1for explanation of observation strategy.Abbreviations in column headings -1 st row: Jm, Jassa marmorata; Bar, barnacle species. 2 nd row: C, colonies; I, individuals; R, recruits; T, tubes.3 rd row: Chg., change in temperature from last observation, except for the 178-day gap ending on April 14, 2014; 13, time period of observations in years (e.g. 13 = 2013).Abbreviations in body of table -g, colonies

Table 4 )
. They did not recruit to the living tissues of itself and other colonial tunicates, solitary tunicates, and the encrusting colonial bryozoan S. unicornis.Growth interactions between recently-settled recruits of D. vexillum and B. violaceus resulted in standoffs.Colonies of D. vexillum were the most aggressive colonizers of the plates in 2014.They successfully overgrew unoccupied plate substrate, recently-settled recruits of D. vexillum, B. violaceus, and S. unicornis, and most, but not all barnacle tests, amphipod tubes, and hydroid stalks (except for some polyps at the ends of the stalks).

Table 5 .
Presence of minor colonizing species.Record of species presence on two 1 × 1 m recruitment plates from September 4, 2013 to March 26, 2015.See the text for explanation of species occurrence, Table1for explanation of observation strategy, and Table4for abbreviations of species names.Abbreviations: Cap, caprellid ampipods; Chg., temperature change; Days, time between observations; Obs., observation number; p, present.