For the time being there are several faunistic schemes of the whole North Polar ocean. They are based on different groups of zoobenthos: Amphipoda (Gurjanova, 1951), Bivalvia (Filatova, 1957), Neptunea (Golikov, 1963), Caprellidae (Vasilenko, 1974), Isopoda (Kusakin, 1979), Buccininae (Golikov, 1980) zoobenthos as a whole (Zenkevitch, 1947). There is no such schemes on Polychaeta.
All mentioned above authors divide the North Polar Basin onto boreal and arctic regions. There is Pacific - Boreal region in schemes of Filatova only, while analogical Atlantic - Boreal region is present in all schemes.
The position of margin between arctic and Atlantic - Boreal region is laid from S-E Greenland to Iceland and further or to the East towards Norway and along its coast up to Kola Peninsula (Zenkevich, 1947; Gurjanova, 1951; Filatova, 1957) or to the North towards Spetsbergen and further towards S-W or S-E coast of the Barents Sea (Golikov, 1963, 1980; Vasilenko, 1974, Kusakin, 1979).
The main difference between schemes are the amount of arctic subregions: from absence of them (Vasilenko, 1974) up to six (Zenkevich, 1947, Filatova, 1957). There are Atlantic and Pacific arctic subregions in four schemes (Zenkevich, 1947; Gurjanova, 1951; Golikov, 1963, 1980), but the position of the board varies among them. Four schemes agree with a presence of deep-water, abyssal (hadal) regions. Filatova (1957) separate bathyal region encountering abyssal one along Eurasia only. There is low-arctic region in schemes of Gurjanova (the White Sea - Svalbad province, including the Bering and White Seas) and Zenkevich and Filatova (the Barents Sea and Chuckchee provinces). Peculiar character of scheme Filatova (1957) and Zenkevitch (1947) is a presence of brackish-water region as a narrow bond along Sibirean coast.
As it is seen from the short review above, zoogeographic scheme differ mainly with amount of regions rather theirs shape. This situation in a great degree is explained with usage if different methods of subdivision. Unfortunately, one cannot understand which method have been used, because they aren't described unsatisfactorily often.
In the present paper we tried to define precision capability of different methods on the base of the same material.
The research is based on new and literature data on a distribution within the North Polar basin of 195 species of Polychaeta belonging to following 18 families: Ampharetidae, Amphictenidae, Capitellidae, Chaetopteridae, Eunicidae, Euphrosinidae, Glyceridae, Goniadidae, Maldanidae, Nephtyidae, Onuphidae (excl. Onuphis s.lato), Oweniidae, Phyllodocidae, Sabellariidae, Sabellidae (excl. Chone), Scalibregmidae, Sternaspidae, Terebellidae (excl. Polycirrinae) (Gorbunov, 1946; Zatsepin, 1948; Annenkova, 1952; Uschakov, 1972; Hlebovich, 1964; Biljard & Carey, 1980; Stop-Bowitz, 1948; Wesenberg-Lund, 1950, 1951, 1953).
We investigated collections of the following scientific cruises: R/V "Sevastopol" (1957-1959) "N.Maslov" (1969, 1970), "Tunetz" (1978), "Alaid" (1980), ice-stations SP-22 and SP-23 (1976-1980) and some samples of others expeditions. This material were collected north to the Chuckchee Sea, Canadian Basin, New Siberia shelf, the Barents, Greenland and Norwegian Seas. The total number of investigated samples is more 1,000; 130 species were identified. The most part of identifications was made by Jirkov, part by A.V. Sikorski, M.V. Kolesnikov, N. E.. Detinova, M. A. Safronova, N. V. Kucheruck. The results of these identifications were partly published (Jirkov, 1980, 1982; Kucheruck, 1980; Tzetlin & al., 1983). Besides I. A. Jirkov more or less through have investigated collections of Zoological Institute in Leningrad (ZIN), Zoological museum of Moscow University, department of invertebrate zoology of Moscow University. N. V. Kucheruck, N. N. Detinova and M. A. Saphronova kindly supply us some of theirs unpublished data.
Authors thank all of them and L. A. Rittikh, L. A. Moskalev, S. S. Drobisheva for supplying by material and V. V. Potin, G. N. Buzhinskaja and V. G. Averintsev for permission to work with collection of ZIN and Ju. E. Romanovskij for consulting in mathematical problems.
As faunistic border we take area of condensing of areal borders, Kuznetzov (1936) named them "synperates". Potentiality of different methods have been determined as possibilities for find these synperates.
In marine zoogeography the most usually used methods are.
1. Taxonomic method. Based on determination of similarity (difference) in taxonomic composition between separate samples or arbitrarily choosing bottom areas. It uses very often for determination of zoogeographical value of boundaries having been found by other methods. We do not use this method, because it is not possible to differ biocenotic boundaries from zoogeographic one using this method.
2. Method of types of areals. The most common and the less precise method - worked in details by Semenov (1972, 1979, 1982). Below we use as names for types of areal term, widely used as names of zones of horizontal and vertical distribution, because there is no principal difference between term describing horizontal and vertical distribution due they determine the boundary of the same area.
3. Method faunistic transects. Based on counting of areal boundaries in equal part of bottom (see Mironov, this volume).
4. Shorygin's method. Based on determination of difference in zoogeographic (not taxonomic) structure of fauna of arbitrarily choosing bottom areas or in samples. Zoogeographic character of analysing areas is proportion of species with different types of distribution. So it is necessary to determine type of areal at first.
We slightly modified Shorygin's method. He supposed that it is enough the simple fact that the number of species with certain type of distribution is more than with other. But there is no reason to suppose, that 50% (when there are two types of distribution) has especial value. We think that it is necessary to determine authenticity. We calculate it as authenticity of difference between proportions of species with different types of distribution within the analysing area and the whole region. We also use for identifying of types of vertical distribution of species the method of depth of mainly occurrence. As it was shown by Zenkevich & Brodskaja (1937) for this purpose biomass is better than density and frequency. So we use biomass when it was possible. But the most species due the low density of populations collected by grab unsatisfactorily, in these cases we use frequency in trawl samples. This method allows to differ species with similar vertical distribution, but having different depth of mainly occurrence. Additional remarks to methods having been used is given in due places.
We suppose possible to analyse distribution of species within the North Polar Basin (from the Bering straight up to the Norwegian and Greenland Seas) only.
We have found seven types of horizontal distribution of Polychaeta within the North Polar Basin (fig. 1-6).
1. Atlantic North Boreal (37 species, fig. 1 a, b): coastal regions of Iceland, Yan-Mayen, Faroes, Shetland, along the West Scandinavia to north up to the North and East Spetsbergen, within the Barents Sea to the line connecting Spitsbergen with southern point of Nova Zemlja, and south-eastern part of the Kara Sea.
2. Pacific North Boreal (9 species, fig. 1c): the most part of the Chuckchee Sea, with exception its east and north parts, also coastal areas of the Beaufort Sea up to Canadian archipelago and probably further.
3. South boreal (48 species, fig. 2): the southernmost parts of the Norwegian and Greenland Seas (coastal regions of West and East Iceland, Faroes, Shetlands and South-West Scandinavia.
4. Atlantic-pacific north boreal (10 species, occurrence of two of them in Pacific sector needs in confirmation, fig. 3): include 1 and 2 types.
5. Arctic-boreal (48 species, fig. 4): the whole the North Polar Basin (there are no limits of distribution within the ocean).
6. Panarctic (17 species, fig. 5): almost the whole the North Polar Basin, with exception of shallow-water regions near South-East Iceland, Faroes, Shetlands and South-West Scandinavia. The main difference from 5 - absence in North-East Atlantic (some species penetrate in the North Sea along the Norwegian trench).
7. High-arctic (7 species, fig. 6): contrary to 6 within the North Polar Basin only, absent in the Norwegian Sea, coastal regions of Iceland, Yan-Mayen, Spitsbergen (but present inside fjords, especially of Spitsbergen), the Beaufort and East and South Chuckchee Sea.
For classification of distribution of the rest 19 (from 195) species there is no enough data.
There is no arctic-boreal, panarctic or high arctic species occurring in west or east parts of the North Polar Basin (these type of distribution were found among other groups of benthic animals).
Synperates divide the North Polar Basin in 4 parts: North Atlantic, high arctic, Atlantic transitional, pacific transitional (fig. 7). The number of regions is smaller than the number of types of areal because different areal include from 1 to 4 region in different combinations. There are no such distinct groups with different vertical distribution.
Formal approach to vertical faunistic transsect was worked by Vinogradova (1958, 1969). She takes as upper and lower bounders of distribution the shallowest and the deepest finds of spices. We prepare four vertical faunistic transsect in regions: 1) between Franz-Joseph-Land and the North Chuckchee Sea; 2) within transitional Atlantic region (see above); 3) within the whole Norwegian Sea; 4) within the East Norwegian Sea (from 65° and 74°N) only. The number of boundaries was calculated for each 100-m horizon.
As the result several synperates were in each transsect, but all of them are artefacts of our different knowledge of fauna of different depth.
The exception is synperate of the last region at the depth of 900-1000 m (fig. 8). This synperate is the result of upper bounders, the absence of 11 species in depth shallower 700 m. But if we take Norwegian Sea as the whole (3 transsect) this synperate disappear almost without a trace.
This synperate allows to divide the Norwegian Sea in shallow and
deep parts.
Used only within transitional Atlantic region (see above) (fig. 7, 9). This region was taken for further investigation, as the only region comparatively evenly covered with available samples.
There are 4 types of areals within the region: Atlantic north boreal, atlantic-pacific north boreal, arctic-boreal and panarctic. First two we consider as one because there is no difference between them within this region.
We agree with Shorygin (1928) than presence arctic-boreal species has less importance for searching of faunistic boundary between arctic and boreal regions than presence of arctic (panarctic) and boreal (north boreal) species. So below we take into account only two types of distribution: panarctic and north boreal (in most cases predominate arctic-boreal species).
In most cases we compare station lists. The exception is deep part of the Norwegian Sea, because of too small number in each sample. So we compare here 5 comparatively large areas: with depth 1400-2000 m through the whole Sea, 2000-3200 and more 3300 in the Norwegian Basin and 2000-2500 and more 2500 m in the Lofoten Basin.
North boreal and panarctic species gives 73 and 27% of total number of species inhabiting this region. We count, that north boreal species predominate if they give more than 73% of total amount species in sample. Taking in account the authenticity of results it is necessary in a case of 15 species, that north boreal give at least 79.5%.
High Boreal region, where north boreal species authentically predominate comparatively small: near Norway to the north up to the south-western Barents Sea. Low Arctic, where panarctic species authentically predominate covers the central part of the Norwegian Sea and the north-east half of the Barents Sea. Transitional zone, where no one group do not authentically predominate as narrow band surrounds the deep part of the Norwegian Sea at south and east, Yan-Mayen and goes to the east up to Kara Sea.
Two parts of Low Arctic region (Norwegian and Barents) can not be differ using Shorygin's method, they are separate only due presence of transitional zone between them. But taxonomic similarity is unexpectedly small.
Regions with predominance are: predominated by north boreal species 80-520 m, predominated by panarctic species 800-3820 m in the Norwegian Sea and 50-390 m in the Barents Sea.
Depth of mainly occurrence of species have determined by graph of biomass (fig. 10) of frequency. Species being characterised by the same depths of mainly occurrence joined in groups. As the result 4 groups were got: species with depth of mainly occurrence on tidal zone, 0-300 m, 500-1500 m and more 2000 m, we call them tidal, shelf, slope and abyssal species. All tidal species are north boreal, all abyssal - panarctic.
It is interesting, that depth of mainly occurrence are almost the same as depth of main geomorphologic structures of the bottom of the North Polar Basin: tidal, shelf, slope and abyssal.
Abyssal species are predominated in deep parts of the Norwegian Sea (more 2000 m) and the Polar Basin (more 1000-1500 m) (fig. 11). Slope species are predominated in narrow band surrounding abyssal one. This legion is laid on depth from 280 (usually from 400 m) up to 1000 m in high arctic and from 200 (usually more 400 m) up to 1500 m in transitional Atlantic region. Station with predominance of shelf species are only in north-west part of the Barents Sea at depth less 300 m.
Taking all schemes (7, 10,11) together we can get following divisions of the North Polar Basin (fig. 12): 1) boreal (North Atlantic), 2) high boreal shelf, 3) low Arctic shelf, 4) low Arctic slope, 5) low Arctic abyssal, 6) high Arctic slope, 7) high Arctic abyssal, 8) Pacific transitional. Transitional areas between these regions are not shown.
Only in transitional Atlantic region there were enough samples for applying Shorygin's method, in others number of samples was less and they placed unevenly, so we did not apply this method (the only exception - separation of high Arctic region to abyssal and slope regions. As the result the different parts of the North Polar Basin were investigated with different precision capability. It is obvious, that after getting additional material Pacific transitional region should be subdivided analogically Atlantic transitional one, shelf regions will appear in all regions shown on fig. 7, the square of slope and abyssal region become bigger, and the position of borders will be corrected.
From the three used methods the less precise one is method of types of areals. Its deficiencies are the result of the absence of adequate classification. The most widely used method "by simple eye" leads to the different results of different authors and to essential distortion of boundaries due generalisation. Tendency to avoid it accepting slightest differences in areals in a fact leads to refusal of classification. For example, Semonov (1979) classified 1500 areals in 561 groups, 314 of them include only species.
The best results method of types of areals gives when big region is investigated, when distortion of areal form are small comparatively to the square getting regions. As artificial approach to minimise distortion is investigation of areals within limited depth. The best results in this case one can get when analysing region can be presented on map as line. The good example are researches of Semenov (1979, 1982) on bottom fauna of South America. But such approach is transitional between method of types of areals and faunistic transsects, because only small parts of faunistic boundaries are under consideration.
Arctic polychaetes by their horizontal distribution (at all depths) have been classified in 7 groups. The coincidence of areal boundaries was so good, that there is no necessity to use method of faunistic transsects. The classification of types of vertical distribution do not gives results.
Method of faunistic transsects does not require classification of areals because it deals with small parts of boundaries. So it is more precise method. Deficiencies of method connected with strong connection of results with knowledge of distribution of species in each small areas. On the other side, when bottom morphology is too complicate it is difficult to impossible to subdivide investigated area in smaller equal ones. So this method is not suitable for investigation of small areas with complicate bottom morphology.
Using this method for investigating of vertical distribution we have got only one faunistic boundary in the East Norwegian Sea. It was possible, probably, due simple bottom morphology and sharp faunistic change in this region.
In others parts of the North Polar Basin shallow and deep water fauna are probably also separated. But the lack of data does not allow us to find faunistic boundary. Within the Barents Sea boundaries of areals and regions are probably too complicate and generalisation leads to missing of information.
Shorygin's method have no both deficiencies: distortion due generalisation have no influence and the results less depend on our knowledge of species distribution. Precision capability is high. So this method is effective in regions with complicate bottom morphology and picture of boundaries.
However it depends on results of classification of areals: the number of regions can not be more than number of areal types. Other deficiency connects with taxonomic diversity of analysing lists (stations): it had to be enough high.
Shorygin's method allows us to get more faunistic regions than others methods. But the connection of boundaries with synperates is not enough clear. In the East Norwegian Sea the boundary getting by Shorygin's method is the same as by method of faunistic transsects, it allows us to suppose that there is good fidelity and boundaries getting by Shorygin's method are possible to accept as faunistic ones.
Transitional zone between boreal and arctic regions getting by
Shorygin's method is wide. Its "arctic" margin has almost
the same position as boundary between boreal and arctic region
of Golikov (1963, 1980) and Kusakin (1979) and "boreal"
one has almost the same position as the same boundary of Zenkevich
(1947), Gurjanova (1951) and Filatova (1957). Analysis of polychaetes
distribution also gives low arctic regions, and besides already
getting by other authors low arctic shelf region we have got also
low arctic slope and abyssal regions. The subdivision of the High
Arctic region in east and west parts is not supported by our results.
[Figure captions]
Puc. 1. Distribution of Asychis biceps (a - our data, á - literature data); Asabellides sbirica. Atlantic and Pacific north boreal types
Puc. 2. Distribution of Lanice conchilega - south boreal type
Puc. 3. Distribution of Polyphisia crassa - Atlantic-pacific north boreal type.
Puc. 4. Distribution of Terebellides stroemi - Arctic-boreal type
Puc. 5. Distribution of Amphicteis ninonae - panarctic type
Puc. 6. Distribution of Melinnexis arctica - high-arctic type
Puc. 7. Division by method of types of areals. a - boreal nothern atlantic, á - transitional regions (Atlantic and pacific), high arctic.
Puc. 8. Distribution along the depth within the East Norwegian Sea. a - total number of species, á - number of upper bounder of distribution, â - number of lower bounders of distribution.
Puc. 9. Division by Shorygin's method.
Puc. 10. Depth of mainly occurrence within transitional region: a - Terebellides stroemi, á - Samythella elongata; â - Melythasides laubieri. Left axis distribution of average biomass, right axis - diapasons of vertical distribution (logarithmic scales).
Puc. 11. Division by method of preferred distribution: a - region of abyssal species predominance, á - region of bathyal species predominance, â - region of sublittoral species predominance, ã - region transitional between 1st and2nd, ä - region transitional between 2nd a and 3rd.
Puc. 12. Synthetic zoogeographic scheme.
[Literature cited not included. Please see original text]