Jump to ContentJump to Main Navigation
Show Summary Details
More options …

Botanica Marina

Editor-in-Chief: Dring, Matthew J.

6 Issues per year

IMPACT FACTOR 2016: 1.239
5-year IMPACT FACTOR: 1.373

CiteScore 2016: 1.28

SCImago Journal Rank (SJR) 2016: 0.456
Source Normalized Impact per Paper (SNIP) 2016: 0.841

See all formats and pricing
More options …
Volume 59, Issue 6


Codium fragile in Norway: subspecies identity and morphology

Caroline S. ArmitageORCID iD: http://orcid.org/0000-0003-2826-5935 / Kjersti Sjøtun
Published Online: 2016-11-17 | DOI: https://doi.org/10.1515/bot-2016-0095


The green alga Codium fragile consists of 10 subspecies, of which subspecies fragile is a well-known invasive seaweed. Morphological work carried out in the 1950s suggested that there were three subspecies along the Norwegian coast: subsp. fragile, subsp. atlanticum and subsp. scandinavicum. However, more recent molecular data have shown the existence of only two subspecies and that these are frequently misidentified. The aims of the present study were therefore to verify which subspecies occur in Norway using the rpl16-rps3 chloroplast marker, to ascertain their likely time of arrival and to compare their morphology to their genetic identity. DNA sequences were obtained for 60 thalli from 18 sites along the coast (57–69° N) and 10 herbarium specimens (1902–1950). The sequences indicated that both subsp. fragile and subsp. atlanticum occur at present and have been in Norway since at least 1932 and 1948, respectively. The subspecies co-occurred at one site, but in general, subsp. atlanticum appears to have a narrower distribution than subsp. fragile, both geographically and in terms of habitat. Importantly, mucron length, other utricle features, or habitat were not always sufficiently reliable to give an accurate subspecies identification, demonstrating the necessity of DNA sequencing for the identification of these subspecies.

This article offers supplementary material which is provided at the end of the article.

Keywords: Codium atlanticum; Codium fragile; herbarium samples; introduced species; morphology


Codium fragile (Bryopsidales, Chlorophyceae) is a siphonous green alga with a NW Pacific origin (Trowbridge 1998). This taxon is presently divided into 10 subspecies (Brodie et al. 2007), one of which is a well-known introduced seaweed: C. fragile subsp. fragile (Suringar) Hariot (previously C. fragile subsp. tomentosoides; Provan et al. 2008). This subspecies has good dispersal and establishment abilities (Nyberg and Wallentinus 2005) and has spread worldwide over the last 200 years (Provan et al. 2008). In new habitats, it can have ecological and economic impacts; for example, it may compete with native kelps or fucoids (Scheibling and Gagnon 2006, Armitage et al. 2014), influence seaweed-associated fauna composition (Schmidt and Scheibling 2006, Drouin et al. 2011, Armitage and Sjøtun 2016), negatively affect commercial bivalve beds (summarised in Trowbridge 1998) and impact ecosystem services (Vilà et al. 2010). In Norway, C. fragile subsp. fragile has been classified as a high-impact non-native species due to its widespread distribution, long expected population lifetime and moderate ecological impact (Gederaas et al. 2012). In some regions, it can become locally abundant, growing in patches in the upper subtidal and infralittoral zones of sheltered and moderately wave-exposed locations, especially with boulder/cobble substratum (Armitage et al. 2014).

According to Silva (1955, 1957), C. fragile subsp. atlanticum (Cotton) Silva and subsp. scandinavicum Silva were already present in Norway in 1946 when subsp. fragile was first recorded, with the first records of subsp. atlanticum from 1895 and subsp. scandinavicum from 1929. These identifications were based on observations of utricle morphology; in particular, utricle dimensions and mucron shape and length have been used to separate the subspecies (e.g. Silva 1955, 1957, Trowbridge and Todd 1999a,b, Brodie et al. 2007). The subsp. scandinavicum was hypothesised to be a northern-adapted subspecies of C. fragile potentially originating from Siberia (Silva 1957), whereas subsp. atlanticum had a more southern distribution, and is listed as observed in Norway, the British Isles, France, the Netherlands, Spain and the Azores (Guiry and Guiry 2015).

However, the use of molecular methods has shown that the status and distribution of the subspecies needs re-examination. Subsp. atlanticum is genetically distinct from subsp. fragile according to a marker in the plastid genome (rpl16-rps3) and has been verified as present in the British Isles (Provan et al. 2008), but there is no genetic confirmation of its distribution in other countries to the authors’ current knowledge. Furthermore, sequences of this marker have revealed the type material of subsp. scandinavicum to be the same as subsp. fragile, uncovering no evidence of the existence of a separate subspecies (Provan et al. 2008). More recently, a comparison of several C. fragile subspecies (including subsp. fragile, but not atlanticum) has suggested that subsp. fragile could be a separate species (Verbruggen et al. 2017).

The introduced subsp. fragile has often been misidentified (Provan et al. 2008). Rojo et al. (2014) found two morphological groups of C. fragile in NW Spain that they initially assigned to subsp. atlanticum and subsp. fragile, but the rpl16-rps3 sequences indicated that all were subsp. fragile. Hubbard and Garbary (2002) and Kusakina et al. (2006) also found morphologically distinct populations of C. fragile in Canada, supported by genetic differences in ISSR nuclear markers (Kusakina et al. 2006), and they suggested that one might be subsp. atlanticum. However, later sequencing of rpl16-rps3 indicated that these were two variants of subsp. fragile (Benton 2014). In Norway, molecular work has been done on only a few samples of C. fragile: Provan et al. (2008) sequenced one sample which was assumed to be subsp. scandinavicum, and Armitage et al. (in press) sequenced 11 samples from the Bergen area, but all of these turned out to be subsp. fragile.

The morphological variability of subsp. fragile and the common misidentification of C. fragile subspecies could imply two things for Norwegian records. The first is that the arrival of the non-native subsp. fragile in Norway could be much earlier than the first record according to Silva (1955, 1957). Observations from the 1930s describe a dramatic and obvious increase of Codium (notes on University Museum of Bergen herbarium specimens include: “found drifting everywhere in great quantities in the sounds in Austevoll”, collected by K. Fægri in 1933; “has spread profusely in Norway in the last few years, earlier nearly unknown here” (translated), collected by H.H.H. Heiberg in 1936). This potentially reflects a rapid expansion of subsp. fragile, years before the first official collection (also see Fægri and Moss 1952). The second is that subsp. atlanticum may not actually be present in Norway, given that misidentification is common and that subsp. atlanticum is only confirmed from the British Isles (Provan et al. 2008).

The aims of this study were therefore (1) to sequence historical collected specimens of C. fragile in order to find the most likely time that subsp. fragile spread to Norway, (2) to check which subspecies of C. fragile are present in Norway, and if more than one is found, (3) to assess whether currently used micro-morphological characters are reliable for their identification. This was done by sequencing the rps3 – rpl16 region of the plastid genome in samples of C. fragile and examining the utricle morphology of the subspecies.

Materials and methods


Fresh samples of C. fragile were collected along the coast of Norway between 57° and 67° N during 20142015 (Table 1). C. fragile has been recorded along the coast of Norway north to around 70° N (Stellander 1969), but is relatively rare in the southeast (Husa et al. 2013), so there were no samples from this area. Two clean branch tips around 3 cm in length from each thallus were dried in silica gel, except when the whole thallus was collected and dried as a herbarium specimen. Samples from Norwegian herbarium collections were also taken (Table 2). Herbaria contacted included the University Museum of Bergen (BG), the Botanical Museum in Oslo (O), the Norwegian University of Science and Technology Museum (TRH) and the Adger Museum of Natural History and Botanical Garden (KMN). Samples were taken of thalli which were identified as subsp. atlanticum in the collections (which looked in reasonable condition), along with other specimens if they were from geographical areas with poor coverage from the fresh material, or from early dates.

Table 1:

Codium fragile: sequenced fresh samples from the coast of Norway (collection locations listed south to north, counties in bold).

Table 2:

Codium fragile: Norwegian herbarium samples from which sequence data could be obtained, listed by date (oldest to most recent) and with herbarium code (Herb. code).

Sequence data

The molecular work in this study was done using methods described by Provan et al. (2008). The primers of Provan et al. (2004) were used to amplify and sequence the rpl16-rps3 region of the plastid genome, which is suitable for indicating evolutionary units within the genus Codium (Verbruggen et al. 2007) and allows identification of subspecies of C. fragile using four single nucleotide polymorphisms (SNPs) (Provan et al. 2008). The UCP6 set encompasses ca. 450 bp, and three sets of primers (UCP61, 2 and 3) divide this up into smaller fragments to allow sequencing of potentially poor-quality herbarium DNA (Provan et al. 2008).

All laboratory work was carried out at the Biodiversity Laboratories (BDL, DNA section) at the University Museum of Bergen/Department of Biology (University of Bergen). DNA was extracted from a small (0.51 cm) section of the dried C. fragile using a Qiagen DNeasy Plant Mini Kit (QIAGEN, Hilden, Germany) according to manufacturer’s instructions. Prior to PCR amplification, the DNA extractions of the fresh samples were diluted by 50, and the herbarium samples by 10. The PCR reaction mix (25 μl total) contained 1 μl 10 μmol forward primer, 1 μl 10 μmol reverse primer, 1 μl DNA, 2 μl dNTP, 2.5 μl 10× PCR buffer, 17.35 μl ddH2O, and 0.15 μl TaKaRa Taq Hot Start version (Takara Bio Inc., Otsu, Japan). The PCR was done under the following thermal settings: initial denaturation at 94°C for 5 min, 5 cycles with denaturation at 94°C for 1 min, annealing at 45°C for 90 s, and extension at 72°C for 90 s, followed by 35 cycles with 94°C for 1 min, 50°C for 90 s and 72°C for 1 min, then a final extension for 5 min at 72°C. Positive and negative controls were routinely used.

Gel electrophoresis was used to check the PCR products. A 1% agarose gel made with 1×TAE buffer (Tris base, acetic acid, EDTA) and containing GelRed (Biotium, Hayward, CA, USA) was loaded with a mix of 4 μl PCR product and 1 μl loading buffer. FastRuler DNA ladder (Thermo Fisher Scientific, Waltham, MA, USA) and images taken with GeneSnap (SynGene, Cambridge, UK) were used to assess DNA size and quantity. PCR products were then purified in 10-μl reactions, containing 8 μl of PCR product, 0.1 μl exonuclease 1 (EXO, 10 U μL−1), 1.0 μl shrimp alkaline phosphatase (SAP 10 U μL−1) and 0.9 μl ddH2O. Incubation at 37°C for 15 min was followed by an inactivation step at 85°C for 15 min. The BigDye (v3.1) method was used to sequence the DNA, using an Applied Biosystems 3730XL Analyzer (Thermo Fisher Scientific) at the Sequencing Facility, Molecular Biology Institute, University of Bergen (Norway).

The programme Geneious (v. 6.1, Biomatters Ltd., Auckland, New Zealand) was used to check sequences, assemble contigs and align the sequences using MUSCLE (multiple sequence comparison by log-expectation). These data were then used to ascertain subspecies identity by comparison with the sequences of Provan et al. (2008) and Benton (2014).

Morphological data

Eleven samples were examined microscopically. These came from six sites: two in Trøndelag (Hellsjæret and Titran) and four in Hordaland (Stora Karlsøy, Bjorøy, Baløy and Tjongspollen; Table 1). Dried tissue was rehydrated in seawater, and utricle morphology was examined approximately 2 cm from branch tips (the area normally used for identification and considered the most consistent; Silva 1957, Dromgoole 1975, Trowbridge 1998). Mucron length and shape was recorded for 16–20 utricles per thallus; these were selected by preparing a slide and measuring the first 20 mucrons that could be clearly seen and were not distorted. Utricles with no mucron at all were not included. Since the starting point for measuring mucron length is not clearly described in every publication, we measured it both from the inner cell wall [hereafter referred to as “length a”, used by Kusakina et al. (2006)] and from the “shoulder” of the utricle (“length b”; see Supplementary Figure S1 for clarification). Utricle shape, length and width, hair scar distance to apex, and whether gametangia were present were also recorded for samples which rehydrated well, for up to 10 utricles per sample. Measurements and images were taken using Leica application suite software (v.4.5), Leica DFC450 camera and DM2000 LED microscope (Leica microsystems, Wetzlar, Germany).


Sequence data

The sequences of the collected samples confirmed that both C. fragile subsp. fragile and subsp. atlanticum currently inhabit the coast of Norway. Although all of the samples from the south, southwest and northern coastlines were subsp. fragile, the islands around Frøya in mid-west Norway support populations of subsp. atlanticum (Figure 1A). This included one sampling site where both of the subspecies were growing together within a radius of around 5 m (site 67, Titran; Table 1).

Codium fragile: Sampling sites along the coast of Norway (excluding northern Norway), showing subspecies identity at each location (subsp. fragile in yellow; subsp. atlanticum in red; uncertain or not C. fragile in grey) according to the rpl16-rps3 genetic marker. Map (A) shows samples from 2014 to 2015, and place names mentioned in the text; map (B) shows herbarium specimens 1902–1950 and their date of collection.
Figure 1:

Codium fragile: Sampling sites along the coast of Norway (excluding northern Norway), showing subspecies identity at each location (subsp. fragile in yellow; subsp. atlanticum in red; uncertain or not C. fragile in grey) according to the rpl16-rps3 genetic marker.

Map (A) shows samples from 2014 to 2015, and place names mentioned in the text; map (B) shows herbarium specimens 1902–1950 and their date of collection.

Sequences of herbarium samples also showed that nearly all were subsp. fragile, including the earliest sequence, which was from a thallus from Hordaland in 1932 (BM5). One thallus, collected from Solund in 1948 (BM12) and originally designated as subsp. scandinavicum by Silva, was genotyped as subsp. atlanticum (Table 2, Figure 1B).

It was not possible to get sequences from a number of the herbarium samples. This included the earliest C. fragile in the collections (1890, Bomlø, Hordaland, B. Hansteen, O) and a floating specimen found on Jan Mayen (1930, J. Lid, O). Short sequences were obtained for the herbarium specimens OM6 and OM7, collected in Ålesund in 1902, but with unexpected results; the sequence for OM7 (136 bp) was most similar to that of Codium vermilara, but had four single nucleotide differences and one 3-nucleotide difference compared to the reference sequences deposited by Verbruggen et al. (2007). The OM6 sequence was short and based on only one strand, which showed double peaks at many of the sites where C. vermilara differs from C. fragile; this may be a result of contamination over the years in the herbarium.

For both subspecies, there is a single-base pair discrepancy between the sequences in the present study and the representative sequences of Provan et al. (2008) to which they were compared (GenBank accession numbers EU045560 for subsp. fragile and EU045559 for subsp. atlanticum); it has been ascertained that these are misreads in the original Provan et al. (2008) sequences (Benton 2014, personal communication J. Provan). Representative sequences from the present study have been uploaded to GenBank; subsp. fragile as accession number KX755326, subsp. atlanticum as accession number KX755327, along with sequences from two additional voucher specimens (whole dried thallus; Table 1).

Morphological data

In the thalli we measured, there were no consistent differences in mucron length between subsp. fragile and subsp. atlanticum (Figure 2). The seven individuals of subsp. fragile had a mean mucron length a of 14 μm (with individual thallus means ranging from 9 to 18 μm), whereas the four atlanticum thalli had a mean of 14 (1318) μm. Using mucron length b also gave similar results for the two subspecies: 19 (1528) μm for subsp. fragile and 18 (1423) μm for subsp. atlanticum. With regard to mucron shape, subsp. atlanticum tended to have fewer pointed mucrons than subsp. fragile, but this character also overlapped between the two subspecies (Figures 3 and 4). In addition, both subspecies could have mucrons with fine striations (Figure 4).

Codium fragile mucron lengths, as measured from the cell wall (length a) and the shoulder of the utricle (length b). Each bar represents the mean mucron length in one thallus, with standard deviation (bars) and maximum lengths (circles) shown (n=16–20 mucrons per thallus). Site number is labelled below the bars, ordered from south to north (Table 1). Subsp. fragile is displayed in yellow, and subsp. atlanticum in red.
Figure 2:

Codium fragile mucron lengths, as measured from the cell wall (length a) and the shoulder of the utricle (length b).

Each bar represents the mean mucron length in one thallus, with standard deviation (bars) and maximum lengths (circles) shown (n=16–20 mucrons per thallus). Site number is labelled below the bars, ordered from south to north (Table 1). Subsp. fragile is displayed in yellow, and subsp. atlanticum in red.

Codium fragile: The frequency of different mucron shapes observed in subsp. fragile (yellow) and subsp. atlanticum (red) thalli (n=16–20 mucrons per thallus). Each bar represents one thallus. Site number is labelled below the bars, ordered from south to north (Table 1). Mucron shapes are clarified in Figure 4.
Figure 3:

Codium fragile: The frequency of different mucron shapes observed in subsp. fragile (yellow) and subsp. atlanticum (red) thalli (n=16–20 mucrons per thallus).

Each bar represents one thallus. Site number is labelled below the bars, ordered from south to north (Table 1). Mucron shapes are clarified in Figure 4.

Codium fragile: Photographs of mucrons and utricles of subsp. atlanticum (A, C and D), and subsp. fragile (B, E, F and G), collected from Titran on Frøya, Norway (63.66618, 8.30521) in November 2014. Mucrons (C) and (G) were categorised as pointed; mucrons like (E) or flatter were categorised as “blunt”; mucrons like (D) and ranging in pointedness towards (F) were categorised as “rounded points”.
Figure 4:

Codium fragile: Photographs of mucrons and utricles of subsp. atlanticum (A, C and D), and subsp. fragile (B, E, F and G), collected from Titran on Frøya, Norway (63.66618, 8.30521) in November 2014.

Mucrons (C) and (G) were categorised as pointed; mucrons like (E) or flatter were categorised as “blunt”; mucrons like (D) and ranging in pointedness towards (F) were categorised as “rounded points”.

Utricle widths were similar between the subspecies, at 271 μm (with individual thallus means ranging from 241309 μm) for subsp. fragile and 274 (248299) μm for subsp. atlanticum. However, the utricles were slightly shorter in subsp. fragile, which had a mean utricle length of 634 (586694) μm whereas atlanticum had a mean length of 711 (680760) μm (Figure 5). The standard deviations of these measurements for each thallus were quite large, indicating much variation. Both subsp. atlanticum and fragile could display a constriction in the middle of the utricle (Figure 4A and B). Gametangia were present in only a few individuals; these were 351 (332374) μm long in subsp. atlanticum (3 thalli, 510 measured per thallus), but only one mature gametangium was seen in the utricles measured for subsp. fragile, which was 256 μm long (see Supplementary Table S1 for full data).

Codium fragile utricle measurements. Each bar represents the mean measurement in one thallus, with standard deviation (for utricle length and width, n=10 utricles per thallus, except for the thallus 67:3, where n=3). “Hair scar distance” refers to the distance of the hair scar from the apex of the utricle (n=1–9 utricles per thallus). For more details of all measurements, see Supplementary Table S1. Site number is labelled below the bars, ordered from south to north (Table 1). Subsp. fragile is displayed in yellow, and subsp. atlanticum in red.
Figure 5:

Codium fragile utricle measurements. Each bar represents the mean measurement in one thallus, with standard deviation (for utricle length and width, n=10 utricles per thallus, except for the thallus 67:3, where n=3). “Hair scar distance” refers to the distance of the hair scar from the apex of the utricle (n=1–9 utricles per thallus). For more details of all measurements, see Supplementary Table S1. Site number is labelled below the bars, ordered from south to north (Table 1). Subsp. fragile is displayed in yellow, and subsp. atlanticum in red.

There appeared to be a difference in the timing of fertility in the area where both subspecies were present. In the thalli sampled from Frøya (sites 67 and 68) in November 2014, all four subsp. atlanticum thalli had mature gametangia, whereas the three subsp. fragile thalli were infertile.


Distribution of Codium fragile subspecies in Norway

By genotyping of herbarium material, the presence of C. fragile subsp. fragile was confirmed for Hordaland by 1932 and for Vest-Adger by 1934, setting back the first collection of this non-native subspecies in Norway by 14 years. However, as these two locations are approximately 300 km apart, it suggests that subsp. fragile had already been spreading for some time before 1932. This is supported by the fact that subsp. fragile had spread through the majority of its current Norwegian distribution by 1950 (approximately 1100 km of coastline), and the fact that subsp. fragile was already present in the Orkneys (N. Scotland) in 1891 (Provan et al. 2008). It seems likely that the dramatic increase in Codium along the Norwegian coast in the early 1930s (Fægri and Moss 1952) was a rapid expansion of C. fragile subsp. fragile, which was also spreading quickly in parts of Ireland during the same decade (see Trowbridge et al. 2013).

In contrast, only a few specimens of subsp. atlanticum were identified in the present study. This subspecies has been present in Ireland since at least 1845 (Provan et al. 2008) and is thought to have spread northwards through Scotland on north-flowing currents (Trowbridge 1998). The distribution in Norway found in the present study is consistent with this mode of dispersal, as the current that flows northwards past the Scotland and the Shetlands tends to hit the Norwegian coast around Stad (Brattegard 2011). This current accounts for a large portion of the marine species which spread naturally to the Norwegian coast (Brattegard 2011). C. fragile is clearly capable of floating long-distances on currents, as shown by the drift specimen found on Jan Mayen (collected in 1930, O), which is approximately 1300 km from the Scottish mainland and 900 km from the nearest point in Norway.

Why subsp. atlanticum has only been found in a fairly limited region of Norway (6164° N) compared to subsp. fragile has a number of possible explanations. It may be that subsp. atlanticum is relatively rare here, and that more samples will reveal a wider distribution. Subsp. atlanticum is also more uncommon and restricted in distribution than subsp. fragile in the British Isles, being absent from areas such as the English Channel (Brodie et al. 2007, Trowbridge and Farnham 2009). Another alternative is that subsp. atlanticum may require higher winter temperatures for survival than subsp. fragile, as seawater temperatures in winter are highest in Norway between Stad and Folla (Brattegard 2011). However, the sample from 1948 just north of Sognefjorden does not fit this pattern. Another possibility is that spread from initial colonisation sites may have been easier for the non-native subsp. fragile than subsp. atlanticum. If subsp. atlanticum spread to mid-Norway from the British Isles, to expand into southern Norway, it would have to disperse against the north-flowing Norwegian coastal current and across several fjord outflows (Bakketeig et al. 2016). On the other hand, some of the earliest findings of subsp. fragile in the present study were from southern Norway; from here, expansion northwards along the coast could be easily achieved by drifting on the coastal current. Subsp. fragile can also be spread by human vectors such as boat traffic (Trowbridge 1998) and can reproduce parthenogenetically and from fragments (Churchill and Moeller 1972, Ramus 1972, Dromgoole 1975, Prince and Trowbridge 2004), meaning a only a small portion of one thallus needs to be transported to start a new population.

In the present study, the subsp. atlanticum samples were from rock-pools on relatively exposed islands, and the shore of a more sheltered location. No subtidal subsp. atlanticum was found. This fits with observations from the British Isles, where subsp. atlanticum has been reported to grow mainly in mid-low intertidal rock-pools (Burrows 1991) or high pools at exposed locations (Trowbridge and Todd 1999a). On the other hand, subsp. fragile can be found in nearly all types of habitat: the subtidal, low intertidal (around or just below mean low water), and rock-pools at both sheltered and more exposed locations. The lack of either subspecies in the mid or high intertidal is unsurprising given that occurrence on emergent substrata is relatively rare for both subspecies in areas where winter freezing occurs (Trowbridge 1998). Because subsp. fragile can be found in a wide range of habitats, we suggest that using tidal position or habitat for subspecies identification can be unreliable (with the possible exception of the subtidal for subsp. fragile, pending further investigation). The observations also suggest that subsp. atlanticum has a more restricted habitat than subsp. fragile in Norway, in addition to the more restricted geographic distribution discussed above. However, sampling at more locations is necessary to confirm this, particularly in mid-Norway and on islands at the outer edge of the coastline. Whether subsp. atlanticum is native or introduced is somewhat uncertain (Trowbridge 1998) but it does not appear to possess the same invasiveness as subsp. fragile on the Norwegian coast.

Morphology of Codium fragile subspecies

The results indicate that utricle morphology is not a particularly reliable character for separating C. fragile subsp. fragile and atlanticum in Norway. There are some trends in the utricle characters which could be related to genetic identity, but there is clearly much overlap and individual variation, making it difficult to use these characters for identification guidelines. Regarding the mucrons, most of the thalli had mean and maximum mucron lengths between 15 and 30 μm, which is intermediate between typical values used for identification of subsp. atlanticum and fragile. Only one extremely long mucron (>40 μm) was seen, and the subsp. fragile mucrons were frequently shorter and blunter than expected (Silva 1957). Thus using commonly applied mucron characters for identification of these thalli would lead to misidentifications; for example that mucrons are <15/20 μm long in subsp. atlanticum (Silva 1957, Burrows 1991, Brodie et al. 2007), mucrons are sharp in subsp. fragile (Silva 1957), and that subsp. fragile has fine concentric striations on the mucrons (Burrows 1991).

Utricle widths were similar between the two subspecies when from the same site, and both could display a constriction (normally only attributed to subsp. fragile; Silva 1957). Distance of the hair scar from the utricle apex was also quite similar: whereas the subsp. atlanticum samples were all within the 130200 μm range described for subsp. atlanticum and below the range of 160260 μm range for subsp. fragile, some of the subsp. fragile samples were also below 160 μm. In addition, although the subsp. atlanticum utricles tended to be slightly longer than the subsp. fragile utricles, they were generally shorter than as described for the subspecies (7801100 μm). Their length and length/width ratio was more typical of subsp. fragile or subsp. “scandinavicum” (5501050 and 480850 μm, respectively; Silva 1957). There were not enough gametangia in the samples to justify a comparison in size or position between the subspecies, and all those observed were either female or, in most cases, indistinct. Determination of the mode of reproduction (parthenogenetic or sexual) has been used as a method of separating subspecies in some studies (e.g. Trowbridge and Todd 1999a) but was not investigated here as most samples were dried and/or without gametangia.

Morphological characters between the diagnostic values for each subspecies are not uncommon in Scandinavian C. fragile, as discussed by Silva (1957). Hybridisation has been proposed as one explanation for “intermediate” morphologies (e.g. Silva 1957, Trowbridge 1998, Kusakina et al. 2006). Theoretically, the two subspecies may be able to hybridise if a male gamete of subsp. atlanticum fused with a female gamete (normally parthenogenetic) of subsp. fragile (Trowbridge 1998). Around Frøya, the subsp. atlanticum thalli were fertile in November whereas the subsp. fragile thalli were not, suggesting reproductive separation in time – but, it is unknown whether an overlap might have occurred before sampling.

However, when considering diagnostic values, it should be taken into account that subsp. “scandinavicum” is likely conspecific with subsp. fragile (Provan et al. 2008). If so, this would mean that the described morphological differences between them may be largely due to environment (as proposed by Fægri and Moss 1952, and discussed in Silva, 1957). If this is the case, it would partially explain Silva’s observation that “intergrades” are fairly common in Norway, lying between the two “morphological plateaus” of subsp. fragile and subsp. “scandinavicum” (Silva 1957), and would imply that the original morphological description of subsp. fragile is too narrow, in particular with regard to mucron length which can be much shorter within individuals of subsp. scandinavicum (described as up to 20 μm; Silva 1957). Some “intermediate” characters between subsp. fragile and subsp. atlanticum may therefore actually lie within the normal range of subsp. fragile, rather than being a product of hybridisation. Large morphological variability in subsp. fragile has also been highlighted in recent work by Armitage et al. (in press) in New Zealand.

The macro-morphology of the two subspecies was not examined here, but some observations were made. It is sometimes stated that the holdfast of subsp. fragile is small (usually <1 cm; Brodie et al. 2007) compared to that of subsp. atlanticum, but personal observations of holdfast size at sites where all sequenced samples were genetically determined as subsp. fragile (e.g. site 18) indicate that the holdfasts can often be much larger than 1 cm in diameter and can spread out in a mossy, undifferentiated way (Supplementary Figure S2). Differences in thallus size and number of dichotomies (e.g. Trowbridge and Todd 1999b) should also be used with caution when the samples are not from the same site. The subsp. fragile found in the present study could occur in patches where all were only around 10 cm long, whereas subsp. atlanticum could be longer than the typical 25 cm (Silva 1957). A difference that may be worth further investigation is that the subsp. atlanticum observed in the present study seemed to have blunter branch tips than subsp. fragile, which had more pointed tips (see Supplementary Figures S2 and S3), but it is unknown if this is influenced by environment.


Both C. fragile subsp. fragile and subsp. atlanticum have been growing in Norway since at least 1932 and 1948, respectively, and are still present today. The distribution of subsp. atlanticum is consistent with spread by currents from the British Isles, and it appears to have a more restricted distribution than subsp. fragile, both geographically and in habitat, but more extensive sampling is needed to confirm this. There are indications of some potential differences between the subspecies in tidal position and timing of fertility, but this also needs further investigation. The results indicate that using micro-morphological or habitat characters to identify subsp. fragile and subsp. atlanticum can easily lead to misidentifications in some locations. Because the morphological characters can overlap between the subspecies, molecular identification is recommended. Genetic identification may allow future studies of the ecology and morphology of these subspecies to reveal further and more reliable differences.


This work was financially supported by a project grant from L. Meltzers Høyskolefond, by the Marine Biodiversity Group at the University of Bergen, and by the Research Council of Norway through the project “Towards integrated European marine research strategy and programmes – SEAS-ERA” (ERAC-CT2009-249552) within the framework of the EU ERA-Net initiative (7th Framework Programme). The study owes much to three groups of people – the first being those who collected samples from along the coast (Vivian Husa, Siri Ødegaard, Mari Heggernes Eilertsen, Annelise Chapman, Ola Vie, Barbro Taraldset Haugland and Katrin Reiss); the second being the staff of the herbaria contacted (BG, O, TRH, KMN); and the third being the staff of the Biodiversity Laboratories at the University Museum of Bergen/University of Bergen, in particular Solveig Thorkildsen and Louise Lindblom. Phoebe Armitage (University of Sydney), Anita Klein (University of New Hampshire) and Jim Provan (Queen’s University Belfast) are also acknowledged for their helpful communication, and the editor and reviewers for their comments.


  • Armitage, C.S. and K. Sjøtun. 2016. Epiphytic macroalgae mediate the impact of a non-native alga on associated fauna. Hydrobiologia 776: 1–15. Google Scholar

  • Armitage, C.S., K. Sjøtun and K. H. Jensen. 2014. Correlative evidence for competition between Fucus serratus and the introduced chlorophyte Codium fragile subsp. fragile on the southwest coast of Norway. Bot. Mar. 57: 85–97. Google Scholar

  • Armitage, P., W. Nelson and J. Sutherland. In press. Mismatch of morphological and molecular identifications in native and invasive subspecies of Codium fragile (Bryopsidales, Chlorophyta). J. PhycolGoogle Scholar

  • Bakketeig, I.E., M. Hauge, C. Kvamme, B.H. Sunnset and K.Ø. Toft (Eds.). 2016. Havforskningsrapporten 2016 [The marine research report 2016]. Fisken og havet, 1–2016. http://www.imr.no/publikasjoner/andre_publikasjoner/havforskningsrapporten/nb-no [Date accessed: 18 August 2016]. 

  • Benton, C.S. 2014. Evaluating and characterizing taxonomic variation, ecology and the transcriptome of Codium fragile in the NW Atlantic (Doctoral thesis). University of New Hampshire. Google Scholar

  • Brattegard, T. 2011. Endringer i norsk marin bunnfauna 19972010 [Changes in Norwegian marine benthic fauna]. Direktoratet for naturforvaltning. Google Scholar

  • Brodie, J., C.A. Maggs and D.M. John. 2007. Green Seaweeds of Britain and Ireland. British Phycological Society. pp. 189–201. Google Scholar

  • Burrows, E.M. 1991. Seaweeds of the British Isles. Volume 2. Chlorophyta. Natural History Museum Publications, London. pp. 190–197. Google Scholar

  • Churchill, A.C. and H.W. Moeller. 1972. Seasonal patterns of reproduction in New York populations of Codium fragile (Sur.) Hariot subsp. tomentosoides (van Goor) Silva. J. Phycol. 8: 147–152. Google Scholar

  • Dromgoole, F.I. 1975. Occurrence of Codium fragile subspecies tomentosoides in New Zealand waters. N. Z. J. Mar. Freshwat. Res. 9: 257–264. Google Scholar

  • Drouin, A., C.W. McKindsey and L.E. Johnson. 2011. Higher abundance and diversity in faunal assemblages with the invasion of Codium fragile ssp. fragile in eelgrass meadows. Mar. Ecol. Prog. Ser. 424: 105–117. Google Scholar

  • Fægri, K. and E. Moss. 1952. On the occurrence of the genus Codium along the Scandinavian Coasts. Blyttia 10: 108–113. Google Scholar

  • Gederaas, L., T.L. Moen, S. Skjelseth and L.-K. Larsen. 2012. Alien species in Norway – the Norwegian Black List 2012. The Norwegian Biodiversity Information Centre. http://www.artsdatabanken.no/Article/Article/133437. [Date accessed: 15 November 2014]. 

  • Guiry, M.D. and G.M. Guiry. 2015. AlgaeBase. National University of Ireland, Galway. http://www.algaebase.org. [Date accessed: 29 April 2015]. 

  • Heggøy, E. 2000. Algevegetasjonen i Tjongspollen, Bømlo, Hordaland [Algal vegetation of Tjonspollen, Bømlo, Hordaland] (MSc thesis). University of Bergen. Google Scholar

  • Hubbard, C.B. and D.J. Garbary. 2002. Morphological variation of Codium fragile (Chlorophyta) in eastern Canada. Bot. Mar. 45: 476–485. Google Scholar

  • Husa, V., A.-L. Agnalt, R. Svensen, K. Rokkan-Iversen, H. Steen, A. Jelmert, E. Farestvedt and H. Pettersen. 2013. Kartlegging av fremmede marine arter i indre og ytre Oslofjord [Mapping of introduced marine species in the inner and outer Oslo fjord]. Utredning for DN 4-2013. Direktoratet for naturforvaltning. Google Scholar

  • Kusakina, J., M. Snyder, D.N. Kristie, and M.J. Dadswell. 2006. Morphological and molecular evidence for multiple invasions of Codium fragile in Atlantic Canada. Bot. Mar. 49: 1–9. Google Scholar

  • Nyberg, C. and I. Wallentinus. 2005. Can species traits be used to predict marine macroalgal introductions? Biol. Invasions 7: 265–279. CrossrefGoogle Scholar

  • Prince, J.S. and C.D. Trowbridge. 2004. Reproduction in the green macroalga Codium (Chlorophyta): characterization of gametes. Bot. Mar. 47: 461–470. Google Scholar

  • Provan, J., S. Murphy, and C.A. Maggs. 2004. Universal plastid primers for Chlorophyta and Rhodophyta. Eur. J. Phycol. 39: 43–50. Google Scholar

  • Provan, J., D. Booth, N.P. Todd, G.E. Beatty, and C.A. Maggs. 2008. Tracking biological invasions in space and time: elucidating the invasive history of the green alga Codium fragile using old DNA. Divers. Distrib. 14: 343–354. Google Scholar

  • Ramus, J. 1972. Differentiation of the green alga Codium fragile. Am. J. Bot. 59: 478–482. Google Scholar

  • Rojo, I., C. Olabarria, M. Santamaria, J. Provan, T. Gallardo, and R.M. Viejo. 2014. Coexistence of congeneric native and invasive species: the case of the green algae Codium spp. in northwestern Spain. Mar. Environ. Res. 101: 135–144. Google Scholar

  • Scheibling, R.E. and P. Gagnon. 2006. Competitive interactions between the invasive green alga Codium fragile ssp. tomentosoides and native canopy-forming seaweeds in Nova Scotia (Canada). Mar. Ecol. Prog. Ser. 325: 1–14. Google Scholar

  • Schmidt, A.L. and R.E. Scheibling. 2006. A comparison of epifauna and epiphytes on native kelps (Laminaria species) and an invasive alga (Codium fragile ssp. tomentosoides) in Nova Scotia, Canada. Bot. Mar. 49: 315–330. Google Scholar

  • Silva, P.C. 1955. The dichotomous species of Codium in Britain. J. Mar. Biol. Assoc. U.K. 34: 565–577. Google Scholar

  • Silva, P.C. 1957. Codium in Scandinavian Waters. Sven. Bot. Tidskr. 51: 117–134. Google Scholar

  • Stellander, O. 1969. Nytt funn av Codium fragile (Sur.) Hariot i Nord-Norge [A new finding of Codium fraile (Sur.) Hariot in Northern Norway]. Blyttia 27: 174–177. Google Scholar

  • Trowbridge, C.D. 1998. Ecology of the green macroalga Codium fragile (Suringar) Hariot 1889: Invasive and non-invasive subspecies. Oceanogr. Mar. Biol. Annu. Rev. 36: 1–64. Google Scholar

  • Trowbridge, C.D. and C.D. Todd. 1999a. The familiar is exotic: I. Codium fragile ssp. atlanticum on Scottish rocky intertidal shores. Bot. J. Scotl. 51: 139–160. Google Scholar

  • Trowbridge, C.D. and C.D. Todd. 1999b. The familiar is exotic: II. Codium fragile ssp. tomentosoides on Scottish rocky intertidal shores. Bot. J. Scotl. 51: 161–179. Google Scholar

  • Trowbridge, C.D. and W.F. Farnham. 2009. Regional comparisons of Codium (Chlorophyta) assemblages in the northern versus southern English Channel. J. Mar. Biol. Assoc. U.K. 89: 255–263. Google Scholar

  • Trowbridge, C., C. Little and P. Stirling. 2013. Temporal changes in Codium spp in Lough Hyne with records for the Southern Irish Coast. Biol. Environ. 113: 1–10. Google Scholar

  • Verbruggen, H., F. Leliaert, C.A. Maggs, S. Shimada, T. Schils, J. Provan, D. Booth, S. Murphy, O. De Clerck, D.S. Littler, M.M. Littler and E. Coppejans. 2007. Species boundaries and phylogenetic relationships within the green algal genus Codium (Bryopsidales) based on plastid DNA sequences. Mol. Phylogen. Evol. 44: 240–254. Google Scholar

  • Verbruggen, H., M.J. Brookes, and J.F. Costa. 2017. DNA barcodes and morphometric data indicate that Codium fragile may consist of two species (Bryopsidales, Chlorophyta). Phycologia. 56: 54–62. Google Scholar

  • Vilà, M., C. Basnou, P. Pyšek, M. Josefsson, P. Genovesi, S. Gollasch, W. Nentwig, S. Olenin, A. Roques, D. Roy and P.E. Hulme. 2010. How well do we understand the impacts of alien species on ecosystem services? A pan-European, cross-taxa assessment. Front. Ecol. Environ. 8: 135–144. Google Scholar

Supplemental Material:

The online version of this article (DOI: 10.1515/bot-2016-0095) offers supplementary material, available to authorised users.

About the article

Caroline S. Armitage

Caroline S. Armitage is a PhD candidate at the University of Bergen, Department of Biology, where she received her MSc in marine biology. Her doctoral research is on non-native seaweed ecology and environmental change.

Kjersti Sjøtun

Kjersti Sjøtun is an Associate Professor at the University of Bergen, Department of Biology. She works on seaweed ecology and biodiversity and is particularly interested in invasive seaweed ecology, population genetics and anthropogenic impact on coastal communities.

Received: 2016-08-22

Accepted: 2016-10-25

Published Online: 2016-11-17

Published in Print: 2016-12-01

Citation Information: Botanica Marina, Volume 59, Issue 6, Pages 439–450, ISSN (Online) 1437-4323, ISSN (Print) 0006-8055, DOI: https://doi.org/10.1515/bot-2016-0095.

Export Citation

©2016 Walter de Gruyter GmbH, Berlin/Boston. Copyright Clearance Center

Supplementary Article Materials

Citing Articles

Here you can find all Crossref-listed publications in which this article is cited. If you would like to receive automatic email messages as soon as this article is cited in other publications, simply activate the “Citation Alert” on the top of this page.

Comments (0)

Please log in or register to comment.
Log in