GEOGRAPHIC AND VERTICAL DISTRIBUTION

Sedimentary vs. water-column materials

As opposed to other zooplanktonic groups, studies on the geographic distribution of extant polycystines have been chiefly based on sedimentary - rather than on planktonic - materials. As mentioned above, sediment samples present some advantages, but also several important shortcomings.

Whereas polycystine abundances seldom exceed five cells per liter in the plankton (e.g., Caron and Swanberg 1990), one gram of (dry) surface sediments can contain thousands to hundreds of thousands of radiolarian skeletons. Plankton samples yield a snapshot-type image of the composition of the assemblages, which does not necessarily adequately reflect long-term trends. The daily, seasonal and interannual variability involved is smoothed out in the sedimentary record, which may be a welcome trait when general patterns are sought. Further, sedimentary materials are more readily available from the various repositories around the globe than plankton samples. In any case, plankton samples not collected for microplanktonic purposes may be useless for radiolarian studies due to inadequate net mesh-size.

On the other hand, interpretation of the geographic distribution of extant radiolarian assemblages on the basis of sediment samples presents several important drawbacks (Boltovskoy 1988, 1994, 1995; Kling and Boltovskoy 1995; Figure 6). On their way to the sea-floor and after settling, radiolarian remains are grazed upon by various consumers thus breaking their skeletons into unidentifiable fragments. Because more delicate shells are destroyed more readily than the more robust ones, specific makeups on the bottom and at mid- depths can differ significantly from the living assemblage in the upper water-column (Boltovskoy et al. 1993b, 1995). Selective dissolution of whole siliceous skeletons en route to the sea-floor and after deposition, although often advocated as an important source of plankton vs. sediments dissimilarities (e.g., Petrushevskaya 1971b; Renz 1976), is probably much less critical than fragmentation due to grazing (Boltovskoy and Alder 1992; Morley et al. Ms). Bottom materials can be reworked after deposition (as a result of which non-Recent deposits, sometimes characteristic of quite dissimilar oceanographic settings, are brought up to the surface layer, or winnowed by bottom currents (dislodging settled skeletons and carrying them thousands of kilometers away; Figure 6). Sediments integrate the imprint of near-surface faunas (which are generally associated with surficial temperature, salinity and primary production fields, as well as with currents and water masses), with the meso- and bathypelagic species whose geographic distribution is uncoupled with upper-water oceanography (Figure 6). In general terms the sedimentary distributions of cold-water species tend to show conspicuous equatorward extensions as compared with their planktonic patterns. This distortion is most probably due to the fact that extended survival of the expatriated cold water taxa is facilitated by submersion (Boltovskoy 1988, 1994; Figure 6); as a consequence, sediment-derived species-specific ranges may wrongly suggest an enhanced tolerance to gradients in the ecological factors.

While these limitations stress the need for caution in biogeographic interpretations based on sedimentary materials, the usefulness of bottom samples for biogeographic, paleobiogeographic and paleoecologic purposes has been confirmed in many reports and is certainly beyond doubt. Furthermore, when compared with water-column materials, sedimentary ones can furnish much useful distributional and ecological data (e.g., Boltovskoy et al. 1993b; 1995), usually unavailable from plankton or sediment trap samples alone.

Geographic patterns

Polycystines are typically open-ocean organisms, occurring throughout the World Ocean. However, distinct coastal associations, while uncommon or absent altogether in areas with an extended shelf, such as the Southwestern Atlantic (Boltovskoy 1980), have been described in various studies. For example, Norwegian fjords host dense and diverse radiolarian assemblages, which differ from those of the open Norwegian Sea (Swanberg and Bjørklund 1986, 1987, 1992). Interestingly, two of these fjord species, Rhizoplegma boreale (probably synonymous with Spongosphaera streptacantha), and Phormacantha hystrix/Plectacantha oikiskos, have been found to strongly dominate (up to 47% of all polycystines) shallow, coastal sediments around Antarctica (Nishimura et al. 1997). General differences between presumably neritic vs. oceanic radiolarian assemblages have been described occasionally in the literature (Kruglikova 1984), and even used for paleoenvironmental reconstructions (e.g., Palmer 1986). However, with the probable exception of specific diversities, which indeed seem lower in neritic assemblages (Nishimura et al. 1997), and the fact that a few selected polycystines are probably less intolerant to near-shore conditions than the bulk, most other traits (such as Spumellaria:Nassellaria proportions, percentages of Spongodiscidae, percentages of "spiny Porodiscidae", percentages of small "Cyrtoidea"; cf. Kruglikova 1984) need further confirmation.

Polycystine densities are typically around 0.3-1 cells per liter, but values exceeding 50 ind. l-1 have been recorded in some productive areas (Caron and Swanberg 1990). The quantitative distribution of polycystines in surface sediments of the South Atlantic is illustrated in Figure 7. This pattern is probably an approximate representation of their concentrations in the water-column as well, and it also roughly reflects the overall distribution of primary production (e.g., Koblentz-Mishke and Vedernikov 1977), and of phytoplanktonic (Semina 1977) and zooplanktonic (Bogorov et al. 1968) biomasses. Highest numbers of polycystines would thus be expected along the upwelling areas off Africa (Abelmann and Gowing 1997), where the highest radiolarian fluxes have been recorded to day (Boltovskoy et al. 1996), and in the equatorial current system. In the southern part of the ocean high densities are probably associated with the subantarctic belt and its northern extensions, the Malvinas (=Falkland) and the Benguela Currents. In a transect between the Antarctic and approximately 30°S, 10°E (off Namibia), Abelmann and Gowing (1997) recorded highest polycystine densities at 100-300 m in Antarctic waters, and at 0-150 m in subantarctic waters (up to 0.3 ind. l-1; these values, however, may be somewhat underestimating, see Boltovskoy and Alder 1992) . In the Southwestern Atlantic (30-60°S, along 55°W), surface (5-15 m) layers were found to host 0.5 polycystines per liter on the average, with maximum concentrations of three shells per liter (Alder et al. 1997). Lowest numbers are those present in Central Gyre and Tropical/Subtropical waters (see Figure 7).

Flux rates of radiolarian shells at depths between 50 and ca. 5000 m vary from 0-4 to over 100,000 ind./m2/day (Boltovskoy et al. 1993a), with highest numbers having so far been recorded in the north-eastern tropical Atlantic (201,064 shells/m2/day; cf. Boltovskoy et al. 1996).

The numbers of species that inhabit the different climatic zones of the World Ocean are difficult to estimate because most authors restrict their scopes to some 20-40 more or less well-defined morphotypes, ignoring the rest of the species. The few surveys that (presumably) did attempt to identify all the skeletons recorded indicate that these numbers oscillate around 100-200 for the tropics and subtropics, dropping to some 50-60 at the poles (Figure 8). This decrease, however, is often punctuated by an isolated peak in the transitional areas which usually host both cold water and warm water taxa, especially in the sediments (see Figure 9) (Boltovskoy 1981d, 1982, 1986).

Despite these rather high numbers, very few of the species are abundant in any given sample. In terms of their relative contribution to the overall polycystine assemblage, usually only one-three species exceed 10%, and up to five represent over 5%; radiolarians whose average percentage abundances are below 1% of the fauna usually comprise 70-90% of all the species recorded (Figure 10). Of the 164 polycystines included in this review, around 10 can attain average proportions in excess of 10% in any given area, 12-15 morphotypes can reach 5-7%, and ca. 50-70 are normally around 1-3% (Table 1). The remaining half of the polycystine species are present at levels below 1%. Highest dominances are associated with polar environments, where a single species or species group can account for 25-40% of the assemblage (e.g., Antarctissa spp. in the Antarctic, cf. Boltovskoy 1987; Amphimelissa setosa in the Greenland Sea, cf. Swanberg and Eide 1992; Phormacantha hystrix/Plectacantha oikiskos and Rhizoplegma boreale, probably synoymous with Spongosphaera streptacantha, in coastal Antarctic sediments, cf. Nishimura et al. 1997; see Figure 10).

Species-specific distributional data for the South Atlantic are scarce and fragmentary. Boltovskoy (1981e) produced a detailed listing of all known Southwestern Atlantic records up to that date, which basically represented 7 reports (Haeckel 1887; Hays 1965; Nigrini 1967; Goll and Bjørklund 1974; Lozano and Hays 1976; Morley 1977; and Boltovskoy and Riedel 1980), chiefly based on sedimentary materials. This objective compilation produced a spotty picture with no discernible patterns. In the 15 years elapsed since that review several contributions based on South Atlantic materials appeared, but they mostly focused on downcore analyses (e.g., Pisias and Moore 1978; Coco 1982; Weaver 1983; Bjørklund and Jansen 1984; Grinstead 1984; Charles and Morley 1988; Alperín 1987), or were restricted geographically to rather small areas (Robson 1983; Dworetzky and Morley 1987; Boltovskoy et al. 1993a, 1993b, 1995, 1996; Abelmann and Gowing 1997). Thus, in order to furnish a more comprehensive insight into polycystine biogeography in the South Atlantic, distributional species-specific data are referred to the 7 distinct areas illustrated in Figure 11). These divisions take into account the distribution of general planktonic biogeographic provinces (e.g., E. Boltovskoy 1970; Koblentz-Mishke and Vedernikov 1977; Boltovskoy 1979, 1981d, 1982, 1986; Dadon and Boltovskoy 1982; Longhurst 1995), as well as radiolarian-based biogeographic patterns (Goll and Bjørklund 1974; Morley 1977; see Figure 11, insets A and B). For some of the especially abundant and better defined taxa relative (percentage) contributions to all polycystines can be predicted with reasonable accuracy. For most others, however, only a very rough indication of their numbers (abundant, present) can be offered for the time being.

The information used to compile Table 1 was not restricted to data from the South Atlantic Ocean, but was extracted from many reports on various oceanic areas, putting special emphasis on water column-based surveys (see "Sedimentary vs. water-column materials" above). Although very subtle differences between oceanic basins probably do exist (Nigrini 1967; Goll and Bjørklund 1974), polycystine species are chiefly restricted in their distribution by climatic and productivity fields, rather than by ocean basins, as are most other pelagic planktonic organisms. Thus, with very few exceptions, similar assemblages characterize the equatorial circumglobal belt, the subtropical zones of the two hemispheres, and the polar waters (Petrushevskaya 1971a). Geographic endemics are rare, probably accounting for less than 5% of all the species (one outstanding example is Antarctissa spp., Figure 15.104, which is absent in the Arctic, but dominates both the plankton and the sediments of the Antarctic zone).

It should be born in mind that the degree of mixture between most of the areas shown in Figure 11 is extremely large. For example, in the western South Atlantic the Transition Zone stretches up to almost 15 degrees in latitude (ca. 34-35°S to 47-48°S; Subantarctic species are regularly found here in the same tows as the Subtropical representatives (E. Boltovskoy 1970, 1981a, 1981b; E. Boltovskoy et al. 1996). Because the Brazil current is a southwest flowing branch of the South Equatorial Current, tropical assemblages differ little from the subtropical ones. Central Gyre fauna is also very similar to the Tropical and Subtropical one, yet these oligotrophic waters, characterized by very low overall plankton abundances, host enhanced proportions of several colonial radiolarians.

Vertical profiles

Vertical profiles of total radiolarian abundance in tropical and subtropical waters indicate that the bulk of their populations is usually located in the upper 50-100 m (Petrushevskaya 1971b; Renz 1976; Dworetzky and Morley 1987; Kling 1979; Kling and Boltovskoy 1995; Abelmann and Gowing 1997; see Figure 12A-B, 12C, 12F-G). Quite often several discrete maxima are recorded, one at or near the surface, and a second one between 50 and 100 m (Petrushevskaya 1971b; Kling and Boltovskoy 1995). In the Antarctic, however, peak abundances seem to be associated with the Warm Deep Water and occur deeper, at 200-400 m (Petrushevskaya 1967; Boltovskoy and Alder 1992; Abelmann and Gowing 1997; Figure 12D, 12E).

Many radiolarian species occupy discrete depth intervals of the water column. Kling and Boltovskoy (1995), on the basis of a series of plankton tows in the upper 2000 m in the eastern subtropical Pacific defined the following characteristic layers: (1) surface (with maxima at 0 m, 25 m, 0 and 50 m, 50 m, or 0 and 100 m), (2) subsurface (maximum at 100 m), (3) deep (maxima at 200 m, 200 and 300 m, or 300 m), and (4) species peaking below 300 m. Roughly similar zonations were established by other authors as well (e.g., Renz 1976; Dworetzky and Morley 1987; Kling 1979). Worldwide depth zonations, however, cannot be defined in terms of fixed depths because the distribution of radiolarian species is related to water masses which move vertically as well as horizontally. For example, in the eastern subtropical Pacific inshore and oceanic 0-25 m waters can host a typically warm-water assemblage associated with the Central Water which is advected coastward by the Southern California Eddy, while midway between these two sites the same depths are inhabited by a conspicuously different, colder-water assemblage associated with the cooler waters of the California Current (Kling and Boltovskoy 1995). Many cold water radiolarians that inhabit the upper layers at high latitudes submerge with their corresponding water masses and can be found at depth in mid- and low-latitude areas (Kling 1976; Boltovskoy 1988; Steineck and Casey 1990). Siphocampe arachnea (Figure 15.167), for example, is a dominant component of surface Pacific Arctic and Subarctic plankton; in the central north Pacific it peaks at 100-300 m, and at 300-1000 m in the subtropical eastern Pacific (Boltovskoy 1994).

Changes in the proportions of presumably living polycystine cells with depth have been assessed in a few studies. Boltovskoy et al. (1993a), based on extensive sediment trap data, concluded that numbers of live specimens decrease drastically downwards (e.g., aprox. 100% at 0 m, 50-60% at 100 m, 20-40% at 200 m, 10-20% at 500 m, 5% at 1000 m; see Figure 12H). These results generally agreee with other studies (e.g., Petrushevskaya 1971a; Kling and Boltovskoy 1995). On the other hand, Abelmann and Gowing (1997) estimated much higher proportions of living cells at comparable levels in the water column: over 90% at 100-200 m, around 70% at 300-500 m. It should be noticed that staining techniques, which are usually applied for these estimates, do not adequately differentiate between live and dead cells (see above "Provenance and collection of materials"), for which reason it is probable that concentrations of living specimens below 50-100 m are systematically overestimated in such surveys (Boltovskoy et al. 1993a; see Figure 12H).

As with geographic patterns, data on the depths at which the various species peak listed in Table 1 have been compiled from reports on different oceanic areas. It is anticipated that they are generally valid for subtropical and tropical environments worldwide; at higher latitudes, however, some deep species may occur closer to the surface, while in the Antarctic the bulk of the asssemblages seems to occupy deeper layers (see above).