Introduction
The genus Globocassidulina belongs to the most common benthic calcareous foraminifera in the western Antarctic Peninsula region inner bay settings (Ishman Reference Ishman1990, Li et al. Reference Li, Yoon and Park2000, Majewski & Anderson Reference Majewski and Anderson2009) as well as on eastern side of the Peninsula (Milam & Anderson Reference Milam and Anderson1981). The species include G. crassa, G. crassa rossensis, (considered by some as an autonomous species G. rossensis), as well as G. biora, and G. subglobosa. The first three are larger and more inflated in shape than G. subglobosa. The last species tends to be rather small in test size, compact, globular, and with a single comma-shaped aperture perpendicular to the basal suture of the final chamber. Globocassidulina crassa, G. rossensis, and G. biora are all closely related and differ primarily by apertural shape. Globocassidulina crassa has a single aperture that is a part of the basal suture of the ultimate chamber, or parallel to it. Globocassidulina rossensis shows a bifurcated aperture, with its branch perpendicular to the basal suture, while G. biora has two apertures parallel to the basal suture of the ultimate chamber (the two apertures while roughly parallel, may join at one end (Quilty Reference Quilty2003)).
There is a considerable degree of inter-gradation between these three types (Finger & Lipps Reference Finger and Lipps1981), with significant ontogenetic changes in morphology of particular species (Nomura Reference Nomura1984), and different understandings of their taxonomic and evolutionary relationships. According to Quilty (Reference Quilty2003), the three morphologically close foraminifera are subspecies of G. crassa. Fillon (Reference Fillon1974) discussed in detail the close relation between G. crassa and G. biora and suggested that the latter is ancestral to the former, as he has not found G. biora among “living” fauna in the Ross Sea. However, G. biora was recorded from Recent sediments around Antarctica by Finger & Lipps (Reference Finger and Lipps1981), Violanti (Reference Violanti1996), Igarashi et al. (Reference Igarashi, Numanami, Tsuchiya and Fukuchi2001), and Majewski (Reference Majewski2005). Fillon (Reference Fillon1974) also suggested that specimens of G. crassa from the Vestfold Hills described by Crespin (Reference Crespin1960) are immature specimens of G. biora that form a continuous range of ontogenetic stages between the immature and mature types. On the other hand, Ward & Webb (Reference Ward and Webb1986) suggested that some G. biora-like specimens of Fillon (Reference Fillon1974, pl. 1, fig. 8) may belong rather to G. crassa and do not show characteristics of immature specimens.
As indicated above, it is rather difficult in practice to unequivocally assign specimens to particular Globocassidulina species, which gives rise to important taxonomic discrepancies between reports from the same regions that apparently discussed very similar fauna. Comparison of different studies on modern foraminifera from bays of King George Island provides an excellent example. Li et al. (Reference Li, Yoon and Park2000) and Yoo et al. (Reference Yoo, Yoon, Park and Kim2006), who analysed Holocene cores from Maxwell Bay, noted the presence of G. biora and G. crassa rossensis. On the other hand, Mayer (Reference Mayer2000) who investigated shallow water assemblages, (down to 30 m), from surface sediment of Potter Cove (part of Maxwell Bay) acknowledged the presence of G. crassa and G. subglobosa. Chang & Yoon (Reference Chang and Yoon1995) noted G. biora, G. crassa, and G. subglobosa in roughly similar numbers from Marian Cove (another part of Maxwell Bay) and water depths less than 100 m. All specimens of Globocassidulina collected by Majewski (Reference Majewski2005) from surface sediments of nearby Admiralty Bay were assigned to G. biora.
Globocassidulina is an important Antarctic genus of foraminifera. According to Ishman (Reference Ishman1990), G. subglobosa and G. biora both dominate biofacies from western Antarctic Peninsula inlets. Li et al. (Reference Li, Yoon and Park2000) noted the association of G. biora with coarse sediment. They speculated that high abundance of this foraminifer may imply a high-energy environment and considered G. biora to be the most important species for palaeoenvironmental reconstructions in shallow water inlet settings due to its high abundance and robustness. This foraminifer was also critical for some palaeoenvironmental reconstructions (e.g. Khim et al. Reference Khim, Yoon, Kim and Shon2001, Yoo et al. Reference Yoo, Yoon, Park and Kim2006, Majewski & Anderson Reference Majewski and Anderson2009). All those facts highlight the need for a better understanding of morphological and molecular variability within the Antarctic Globocassidulina.
Material and methods
Surface sediment samples for population and morphological study of Globocassidulina were taken at 21 locations in Admiralty Bay from water depths between 8 and 254 m in early 2007 (Table I, Fig. 1). They were wet sieved through 63 μm and 125 μm sieves and stained with Rose Bengal (1g l-1) and 70% ethanol diluted in seawater. The stained residues were washed in tap water and dried. In most samples, all “living” (stained) foraminifera were picked from the 63–125 μm and >125 μm fractions. Few faunally-rich samples were divided using a dry microsplitter. The maximum test diameter of all stained Globocassidulina from these samples was measured optically at 63x magnification with ±16 μm precision. Apertural shapes of selected specimens were analysed under SEM.
List of sampling stations indicating site location, its water depth, and method of sampling.
The location of a. King George Island, b. Admiralty Bay, and c. sampling stations (KG1 through KG24) with G. biora abundances shown by circles expressing number of specimens per cm2, and percentages in total Rose Bengal stained assemblage. Stations with the percentages only were sampled by scuba divers. Dark gray indicates inland glacier free areas (Battke Reference Battke1990). The bathymetric contour lines in metres after Straten (unpublished).
Nineteen specimens of Globocassidulina representing different morphological types were fixed for DNA examination (Table II). They were isolated from the upper 2–3 cm of surface sediments collected at seven locations. As soon as possible after collection, the sediment was washed with cold seawater and wet sieved through 125 μm sieves. The residues were stored in the refrigerator at about 2°C, for periods of several days. The foraminifera were sorted under a binocular microscope from the residues kept cool on a tray of ice. Specimens for molecular study were transferred one by one into guanidine or AP1 (DNAEasy, Qiagen) extraction buffers.
List of DNA samples used in this study.
A 3′ fragment of the SSU rRNA gene was successfully amplified for 17 DNA isolates (specimens). Foraminiferal specific primer s14F3 and universal eukaryotic primer sB were used for the first amplification. Foraminiferal specific primer s14F1 and primer sB were used for reamplification. PCR conditions and primer sequences are described in Pawlowski (Reference Pawlowski2000).
The ITS region (ITS1 + 5.8S + ITS2) was successfully amplified for 18 DNA isolates. The PCR amplification was performed using the foraminiferal-specific primer pair s14F3 and 2TAIC and re-amplified with the universal eukaryotic primer sBr and foraminiferal specific primer 2TAIC. The compatibility of the SSU and ITS sequences was tested by amplifying a longer fragment, using primers s14F1 and For2 situated in the central part of ITS region (Fig. 2). Primers used for amplification of ITS region are described in Pawlowski et al. (Reference Pawlowski, Fahrni, Lecroq, Longet, Cornelius, Excoffier, Cedhagen and Gooday2007).
Schema of rDNA showing the position of primers used for amplification of SSU and ITS rDNA (dark and bright gray arrows, respectively) and for testing the compatibility of the SSU and ITS sequences (black arrows).
The amplified PCR products were purified using High Pure PCR Purification Kit (Roche Diagnostics). They were ligated in the Topo Cloning vector (Invitrogen), and cloned using ultracompetent cells XL2-Blue MRF’ (Stratagene). Sequencing reactions were prepared using an ABI-PRISM Big Dye Terminator Cycle Sequencing Kit and analysed with an ABI-PRISM 3100 sequencer (Applied Biosystems), all according to the manufacturer’s instructions.
In the SSU analysis, 31 sequences obtained after cloning of 17 PCR products were manually aligned using Seaview software (Galtier et al. Reference Galtier, Gouy and Gautier1996) to 21 sequences of foraminifera from Cassidulinidae family, obtained mostly from the same region. Highly variable regions of the alignment were removed, leaving 1163 sites for phylogenetic analysis. The phylogenetic tree was constructed using maximum likelihood method with GTR + GI model, using Treefinder (Jobb et al. Reference Jobb, von Haeseler and Strimmer2004). The same programme was used to analyse 34 sequences of the ITS region.
Results
Morphologic diversity of Globocassidulina in Admiralty Bay
Among the stained samples from the recent survey in Admiralty Bay, Globocassidulina was found everywhere except at station KG15 located within Martel Inlet, next to the Keller Peninsula (Table I, Fig. 1). In eight samples, substantial populations (>50 specimens) of stained Globocassidulina were encountered. Among truly living specimens, recognized by coloured cytoplasm rather than by Rose Bengal staining, four morphological types were identified. They include 1) large inflated forms with broad apertural area, doubled aperture and pink to yellow cytoplasm, 2) large forms with more globular outline, doubled aperture and pink to green cytoplasm, 3) medium in size with bifurcated or double aperture and greenish cytoplasm, and 4) small forms with single aperture and greenish or yellowish cytoplasm. The presence of four morphological types among Globocassidulina in Admiralty Bay was supported by SEM inspections of stained specimens (Fig. 3). Their characteristic features are summarized in Table III.
SEM images of Globocassulina from Admiralty Bay, belonging to the four morphologic types. Type 1 (large flat): 1. KG7; type 2 (large globular): 2. KG8; type 3 (middle in size): 3. KG7, 4–6. KG8; type 4 (G. subglobosa-like): 7. KG20, 8. KG23.
Test size measurements (Fig. 4) on the total of 1277 specimens confirmed the presence of considerable morphological heterogeneity in Globocassidulina population. In the shallowest sample with substantial population of stained specimens (KG7, 30 m depth), 180 specimens show bimodal test size distribution. Most specimens have tests of 600–750 μm, while the rest are smaller than 500 μm. The group of samples from water depths between 40 and 179 m (KG8, KG9, KG10, KG13, KG14, KG19, and KG24) show a more homogenous test size distribution with only a few specimens larger than 600 μm and the overwhelming majority within a broad range between 100 and 600 μm, but rather smaller than 400 μm. In contrast, the deepest water samples from 223 to 254 m (KG20, KG21, KG22, KG23) show the greatest numbers of specimens between 100 and 200 μm, and significantly fewer specimens that are up to 500 μm in test size. To summarize, the presence of at least three morphological types within Globocassidulina population in Admiralty Bay, is generally confirmed by the test size results.
Test size (maximum length) measurements from 12 locations, arranged according to increasing water depth. Note different scale on the composite, most right graph.
Genetic diversity of Globocassidulina in Admiralty Bay
To verify whether observed morphological differences correspond to genetic differentiation between the morphological types, we amplified and sequenced a 3′ fragment of the SSU and the complete ITS region of the rDNA of specimens representing all morphotypes. The analysis of SSU rDNA (Fig. 5) shows considerable homogeneity between specimens, except for the isolate 8234, which may represent G. subglobosa. All morphologically distinctive types of Globocassidulina from Admiralty Bay merge into one group clearly distinctive from other Cassidulinidae. Because of the presence of a doubled aperture in the largest, most characteristic specimens, we consider this group to represent G. biora. Moreover, four isolates suspected of belonging to G. subglobosa (7995, 8026, 8087, 8088) blend within the G. biora cluster, suggesting that they may be in fact immature G. biora.
SSU rDNA tree showing relations between selected Cassidulinidae from various high-latitude locations.
The test whether the genetic homogeneity of the SSU rDNA sequences is not due to the relatively low mutation rate in this gene (Bowser et al. Reference Bowser, Habura and Pawlowski2006), we also sequenced more rapidly evolving ITS rDNA region. Phylogenetic analysis of ITS sequences gives a tree similar to the one obtained using SSU rDNA, with no clear differentiation between different morphotypes, except for the DNA isolate 8234, probably representing G. subglobosa (Fig. 6). Within the G. biora group, the distances between sequences obtained from different isolates (inter-isolate variation) averaged 0.49%, with the range between 0 and 1.21, while between the two sequences obtained from the isolate 8234 and G. biora the distances were almost an order of magnitude larger and averaged 3.9% with the range between 3.59 and 4.40%. The intra-isolate variation was an average of 0.60%, actually exceeding the inter-isolate variation within the G. biora cluster. To conclude, comparison of the ITS sequences obtained from the four morphological types within G. biora falls within 0.38–0.60% range and reveals no significant differences between intra- and inter-morphological variability (Table IV), confirming the monospecific character of the G. biora cluster.
ITS rDNA tree showing sequences obtained from different morphotypes of G. biora.
Average sequence divergence (in percentage) between different morphological types of G. biora from Admiralty Bay, based on the analysis of the ITS fragment of the rRNA gene.
Measurement of the proloculus
For analysing sizes of proloculi, 32 specimens of Globocassidulina from Admiralty Bay were X-rayed (Fig. 7). In the case of 28 micrographs, it was possible to measure the sizes of proloculi. In the larger specimens (type 1 and 2) a degree of uncertainty in identification of proloculi existed. However, dissection of 30 specimens from KG7 and KG8 confirmed presence of large proloculi (∼120 μm) in majority of specimens. Only three dissected specimens were found to be microspheric. The plot of the proloculi sizes against test diameters (Fig. 8) shows no significant difference in the size of proloculus between types 1 and 2. Their average diameter of ∼140 μm exceeds values measured on dissected specimens, due to rather obscured X-ray view of inner whorls in the larger and thicker-walled specimens. Type 3 shows proloculi at ∼125 mm in average, which stays in agreement with the values measured on the dissected specimens of types 1 and 2. On contrary, the smallest specimens, represented by type 4, show relatively small proloculi (∼80 μm on average or less). All specimens of type 4, that were X-rayed, were from samples 20 and 23 (249 and 223 m) and they may represent G. subglobosa, similar to the isolate 8234 that was the only one from below 200 m analysed genetically in this study and which SSU and ITS sequences clearly differed from remaining Globocassidulina isolates (Figs 5 & 6).
X-ray images of four morphological types of Globocassidulina from Admiralty Bay. 1. Type 1, KG7, 2. Type 2, KG8, 3 & 4. Type 3, KG8, KG14, 5 & 6. Type 4, KG20, KG20.
Test versus proloculus diameter among morphologic types of Rose Bengal stained Globocassidulina in Admiralty Bay. Note that average proloculus diameter measured on dissected specimens from type 1 and 2 was on average ∼120 μm. The difference is due to obscured X-ray views of inner whorls in the larger and thicker-walled specimens.
Discussion
What is the reason for morphological diversity of G. biora?
Despite the presence of considerable morphological variability within Globocassidulina in Admiralty Bay (Table III, Fig. 3), the molecular analysis of the SSU and ITS fragments of rRNA gene indicates that except for minute forms from deeper than 200 m probably representing G. subglobosa, the rest of Globocassidulina from Admiralty Bay is monospecific and belongs to G. biora (Table IV, Figs 5 & 6). What could be the reason for the significant morphologic variability within G. biora? There appears to be a few possibilities: 1) development of distinctive micro and megalospheric generations, 2) a presence of distinct cohorts that at the time of collection were at different growth stages, and/or 3) it could be due to environmental variability in benthic habitats throughout Admiralty Bay that resulted in development of different morphologies.
The first possibility was tested by analysing inner-test morphology of the four types of G. biora. Although the life cycle is more varied in foraminifera than in any other group of protists (Goldstein Reference Goldstein1999) and no specific information on life cycle in Cassidulinidae is available, in general, the foraminiferal life cycle is characterized by alternation of sexual and asexual generations. As a result, megalospheric gamonts with a large proloculus and relatively small overall test diameter, and microspheric agamonts with tiny proloculus but relatively large overall test may develop (Goldstein Reference Goldstein1999). Nomura (Reference Nomura1984) reported the presence of megalospheric and microspheric forms in G. biora from raised beaches and modern deposits of the Lützow–Holm Bay, East Antarctica. He reported the presence of two types of tests having a proloculus of 72 μm in average and of less than 10 μm in diameter.
It appears that among G. biora, morphologic dimorphism, i.e. a distinction between micro and megalospheric generations, is not responsible for the morphological variability observed. However, clear difference in the size of proloculus is observed between G. biora and G. subglobosa (∼120 μm vs ∼80 μm; Fig. 8). Moreover, megalospheric proloculi of G. biora from Admiralty Bay (∼120 μm) are on average larger than those reported from Lützow–Holm by Nomura (Reference Nomura1984) (72 μm).
The second possibility of significant morphological changes throughout ontogeny among Globocassidulina was noted by Finger & Lipps (Reference Finger and Lipps1981). They suggested that smaller specimens of G. crassa rossensis and G. biora are similar to G. subglobosa, but grade into larger and more compressed forms with branched or doubled apertures. In the same study, they reported dissection revealing that earlier chambers of adult G. crassa rossensis and G. biora display fully developed apertures without any indications of apertural growth stages. Thus, their final conclusion was not unequivocal. Our data suggest that a large part of morphological variability observed among G. biora in Admiralty Bay may be a result of observing specimens at different growth stages. The presence of single, bifurcated, and doubled apertures among the four morphological types that, according to the molecular analysis all fell into the G. biora cluster, suggests that throughout ontogeny the aperture becomes first bifurcated and then matures into doubled. Similarly, the test shape becomes more inflated in the largest specimens, as observed in type 1 (Fig. 3). All these findings are in agreement with the conclusions of Nomura (Reference Nomura1984) who presented analogous morphological variability through ontogenic development of G. biora from the Lützow–Holm Bay.
On the other hand, the ontogenic changes cannot explain morphological differences between type 1 and type 2, both representing relatively large specimens (Table III, Fig. 3) and falling within a single group on the proloculus vs test size plot (Fig. 8). Dissection of 30 specimens revealed that in both types the megalospheric forms dominate; however, a few microspheric forms are also present. Significant intraspecific morphologic variability is not unknown among benthic foraminifera. Study of Ammonia showed wide morphologic diversity in natural habitats and cultured populations. Up to six different “morphospecies” were found in their natural habitats and in clonal cultures that were suggested to be ecotypes of Ammonia beccarii developed due to environmental factors (Schnitker Reference Schnitker1974).
A review of ecological parameters that may have impact on benthic foraminiferal tests morphology was presented by Boltovskoy et al. (Reference Boltovskoy, Scott and Medioli1991). It appears that the differences in test shape between types 1 and 2 of G. biora may be also a consequence of ecological variability in benthic habitats throughout Admiralty Bay. A problem with this interpretation however, is that the large globular specimens of morphological type 2, that are encountered at several locations (KG5, KG8, KG9, KG13, KG19, KG24), at all locations except KG5 and KG24 blend with the inflated specimens of type 1. Therefore it is difficult to propose any hydrographic (e.g. temperature, salinity, calcium carbonate solubility etc.) or substrate-related factor that could lead to development of two different morphologies in G. biora at the same locations. Among the ecological parameters, reviewed by Boltovskoy et al. (Reference Boltovskoy, Scott and Medioli1991), only the nutrition, or more precisely food type could be a candidate for the development of the two different types of large G. biora. There are numerous studies showing morphological and cytological response of benthic foraminifera to heavy metal pollution (e.g. Alve Reference Alve1991, Le Cadre & Debenay Reference Le Cadre and Debenay2006) and it was suggested that the pollutants may penetrate the foraminiferal cell together with food (Yanko et al. Reference Yanko, Ahmad and Kaminski1998). However, so far no data show that different types of food could actually lead to development of different test morphologies in benthic foraminifera. In the case of G. biora from Admiralty Bay, the difference in cytoplasm colouration between types 1 and 2 (Table III) may give slight support for such possibility.
Is there only one large Globocassidulina (G. biora) in the South Shetland Islands and Antarctica?
The first detailed study on Recent benthic foraminiferal distribution in Admiralty Bay in the 2002/03 summer season used only the > 125 μm fraction (Majewski Reference Majewski2005). Specimens of Globocassidulina were encountered at all locations, between 8 and 520 m, except for a single site in Hervé Cove (8 m), where only a few Quinqueloculina were found. All relatively large specimens of Globocassidulina were reported as G. biora. However, a considerable variability in their apertural shapes was noted. It was also clear that although G. biora inhabits practically the entire Admiralty Bay down to at least 520 m, it dominates shallow water assemblages at less than 200 m (Majewski Reference Majewski2005). However, other studies in South Shetland Islands (Finger & Lipps Reference Finger and Lipps1981, Chang & Yoon Reference Chang and Yoon1995, Mayer Reference Mayer2000, Li et al. Reference Li, Yoon and Park2000, Gray et al. Reference Gray, Sturz, Bruns, Marzan, Dougherty, Law, Brackett and Marcou2003, Yoo et al. Reference Yoo, Yoon, Park and Kim2006), have reported the presence of more than one representatives/species of Globocassidulina. Keeping this fact in mind, along with an investigation of living monothalamous foraminiferal assemblages in Admiralty Bay (Majewski et al. Reference Majewski, Lecroq, Sinniger and Pawłowski2007), particular attention was paid also to morphological diversity among living Globocassidulina.
The molecular homogeneity, despite considerable morphological variation suggests that G. crassa reported from modern sediments at King George Island by earlier studies (e.g. Chang & Yoon Reference Chang and Yoon1995, Mayer Reference Mayer2000) may be in fact G. biora as might G. crassa rossensis reported from Holocene sections by Li et al. (Reference Li, Yoon and Park2000) and Yoo et al. (Reference Yoo, Yoon, Park and Kim2006). In studies of recent Antarctic benthic foraminifera, Finger & Lipps (Reference Finger and Lipps1981) have already suggested a common over-reporting of the true G. crassa type of d’Orbigny (1839). They suggested that many workers have ignored the diagnostic apertural branches that differentiate other taxa from G. crassa, having only a single aperture. Interestingly, despite the comment of Finger & Lipps (Reference Finger and Lipps1981), Gray et al. (Reference Gray, Sturz, Bruns, Marzan, Dougherty, Law, Brackett and Marcou2003) who subsequently studied recent deposits at the same locations, listed G. crassa along with G. biora as present at Deception Island, even though their specimen illustrated as G. crassa seems to have a doubled aperture. They did, however, point out difficulties in classifying Globocassidulina in cases of poor aperture preservation.
Finger & Lipps (Reference Finger and Lipps1981) acknowledged the presence of two large Globocassidulina taxa in Deception Island (∼150 km south-west from King George Island). They dissected some large specimens of G. biora and what they regarded as G. crassa rossensis and noted that their inner chambers displayed fully developed apertures without any indication of apertural growth stages. Their observation was questioned by Nomura (Reference Nomura1984), who observed changes in apertural shape from single, through L-shaped into double, parallel in mature stage of G. biora, in agreement with another suggestion of Finger & Lipps (Reference Finger and Lipps1981) about potential difficulties in discriminating between immature specimens of G. biora and G. crassa rossensis. In the view of the present findings, it seems to be likely that all specimens reported from Deception Island may in fact belong to G. biora.
It appears that at least some of the morphological types of G. biora distinguished in Admiralty Bay may have broad pan-Antarctic distribution. Crespin (Reference Crespin1960) who described G. biora from Vestfold Hills illustrated large, globular specimens that could correspond to type 2 from this study. Similarly, large specimens from raised beaches of Lützow–Holm Bay (Nomura Reference Nomura1983, Reference Nomura1984) and from the Firth of Tay Holocene section (Majewski & Anderson Reference Majewski and Anderson2009) resemble type 2. On the other hand, in the Ross Sea, Violanti (Reference Violanti1996) reported forms that may correspond to type 1 in this study. She noted G. subglobosa, G. crassa, G. rossensis, and G. biora in a single nearshore sample (IB 3) from near Terra Nova Station (Ross Sea) and noted a more compressed shape of the largest G. biora as compared with G. crassa. A large and inflated specimen of G. biora was illustrated by Ward & Webb (Reference Ward and Webb1986) from materials deposited onshore Ross Island by the ice shelf. It seems to correspond to the type 1 from Admiralty Bay as well. The living specimens of type 1 morphotype were also observed by one of us (JP) at shallow depths in McMurdo Sound.
According to Quilty (Reference Quilty2003), the three morphologically close species (i.e. G. crassa, G. rossensis, and G. biora) are in fact subspecies of G. crassa. Their very close phylogenetic relationship is in concordance with the ontogenetic development of the aperture in G. biora (Nomura Reference Nomura1984) and also reflected in the molecular homogeneity among forms with differently developed apertures in Admiralty Bay. The absence of the true G. crassa in Deception Island and possibly at other Antarctic locations, as suggested by Finger & Lipps (Reference Finger and Lipps1981), may imply that G. crassa, originally reported north of the Polar Front, and G. biora, common in recent Antarctic settings, could represent two distinct species with populations restricted to the Magellan region and the Antarctic, respectively. It seems probable that large, typical specimens of G. biora are more representative for shallow waters (Bernhard Reference Bernhard1987, Li et al. Reference Li, Yoon and Park2000, Majewski Reference Majewski2005) than for greater water depths (Fillon Reference Fillon1974, Osterman & Kellogg Reference Osterman and Kellogg1979, Kellogg & Kellogg Reference Kellogg and Kellogg1987), mainly due to control by water mass properties with respect to CaCO3 dissolution (e.g. Anderson Reference Anderson1975, Ishman & Szymcek Reference Ishman and Szymcek2003). This bathymetric constraint could limit Globocassidulina gene flow between the Antarctic and the Magellan region.
Conclusions
Four distinctive morphological types found among living Globocassidulina in Admiralty Bay include 1) large inflated forms with a broad apertural area, doubled aperture and pink to yellow cytoplasm, 2) large forms with a more globular outline, doubled aperture and pink to green cytoplasm, 3) medium forms with a bifurcated or double aperture and greenish cytoplasm, and 4) small forms with a single aperture with greenish or yellowish cytoplasm. The molecular analysis of the SSU and ITS rDNA indicates that they are monospecific and belong to G. biora. Among the smallest forms, only a single isolate, the only one from deeper than 200 m, stands out in the molecular analyses. It may represent the true G. subglobosa. The morphological types of G. biora seem to represent populations at different stages of ontogenetic development. However, the variability among large G. biora collected at the same locations is difficult to interpret. It seems that different types of food may be the most likely factor that could lead to development of different test morphologies in adult G. biora. It also seems probable that G. biora is the only recent, large, shallow water Globocassidulina represented throughout South Shetland Islands and possibly throughout the Antarctica, while G. crassa is typical of the Magellan region north of the Polar Front.
Acknowledgements
This study was supported by a grant of Polish Ministry of Science and Higher Education No N307 115135, a grant of Swiss National Science Foundation 31003A_125372, and the G. & A. Claraz Donation. We would like to thank Professor Johann Hohenegger (Vienna University) for his kindest help in obtaining X-ray micrographs, Dr José Fahrni and Jackie Guaird for technical assistance, and Drs Béatrice Lecroq and Frédéric Sinniger for help in collecting material. We would also like to thank Alessandra Asioli and an anonymous reviewer for useful suggestions for improvement to this manuscript.
