On the biogeography of Cumacea ( Crustacea , Malacostraca ) . A comparison between South America , the Subantarctic Islands and Antarctica : present state of the art *

Cumacea (Crustacea) were collected during the “Joint Magellan” expedition in November 1994, by means of an epibenthic sledge from RV “Victor Hensen”. The cumaceans were well represented, the second abundant order after the amphipods, among the other Peracarida in depth ranges between 25 and 665 m. Twenty-five species were found in the samples mainly from the Beagle Channel, nine of them were already known for this region. 14 species were recorded for the first time for this region, 2 of them were known from the northern Argentinian coast and one from Antarctica. The most important in terms of species richness and abundance were the families Diastylidae, Nannastacidae and Leuconidae. In the Beagle Channel an almost completely different cumacean fauna was found compared to the Subantarctic Islands, the Antarctic Peninsula and eastern Antarctic (Prydz Bay) regions. Comparison of published data and the present results show moderate overlap in the cumacean fauna at the species level between the periantarctic South Georgian shelf / Antarctic Peninsula (48%). Little correspondence at the species level was found between Antarctica / Subantarctic Kerguelen (14 %), South Georgia / Kerguelen (13 %) and Magellan / Antarctica (11 %). Interestingly, the Magellan region and South Georgia show very little species overlap (5 %). It is concluded that the Antarctic shelf regions were not colonized from the Magellan region via the Scotia Arc.


INTRODUCTION
Very little is known about the fauna of the southern tip of South America and some parts of the Antarctic in general.This is especially true for the cumaceans of the Magellan region, as there have been few benthos expeditions into this part of the world: The "Hamburgische Magellanische Sammelreise 1892/93" (Zimmer, 1902;Panning, 1957); the "Vema" expedition 1958-1961, Lamont Geological Observatory, USA, and the study from the "Joint Chilean-Italian-German Victor Hensen Campaign, 1994" (Arntz and Gorny, 1996).
In the present study the Magellan region is defined as the shelf of the South American continent south of 40°S including the Falkland Islands.
In 1902 Zimmer described five new cumacean species.Eight species were known at that time for the Magellan region.Until the early 1990s, after Bacescu and his collegues (Bacescu and Muradian, 1974;Muradian, 1976;Petrescu, 1995) examined the "Vema"-material, 26 species were reported for the Magellan region.The present "Victor Hensen" expedition increases our knowledge on the Cumacea from this region.The Antarctic Cumacea are much better known than the Magellan species, due to the results of various international Antarctic expeditions which were undertaken since the 1950s (Ledoyer, 1993).
The major aims of the present study were: 1. to describe the composition and diversity of the Cumacea fauna in the Magellan region; 2. to examine the similarity, species overlap and endemicity of the Cumacea fauna from the Magellan and the Antarctic region; 3. to discuss the question of faunal distribution in order to scrutinize whether the Antarctic was colonized from South America via the Scotia Arc.
The analyses of these aspects should help us to gain a better understanding of the interrelationship of the cumacean fauna between these regions.Published data and the present study on the Cumacea of the Magellan region are compared with the available information from Antarctica.The Antarctic is defined according to the definition of Hedgpeth (1969), including the shelf of South Georgia Island.

MATERIAL AND METHODS
Cumacea were collected during the "Joint Chilean-Italian-German Victor Hensen Campaign, 1994" on board the research vessel "Victor Hensen".The largest amount of Cumacea was obtained during daytime by an epibenthic sledge (Brandt and Barthel, 1995).The results are focussed on these samples.A smaller number of Cumacea was collected by Rauschert's small dredge (opening: 0.  equipped with a net of fine (0.3 mm) mesh size to sample smaller crustaceans.The geographical location and sampling depths of the epibenthic sledge stations are listed in detail in Table 1, for details of the dredge stations see Arntz and Gorny (1996).
A list of all Cumacea species known from the entire study region is presented herein in order to facilitate a comparison of the distribution of cumaceans from the Magellan and Antarctic regions.This list is based on Bacescu's account ( 1988,1992) and includes those studies published later by Petrescu (1991Petrescu ( , 1994Petrescu ( , 1995)), Ledoyer (1993) and Mühlenhardt-Siegel (1994, 1996).

RESULTS
In the present investigation 25 species belonging to five families were identified from the Magellan region.Five of them are new to science.Two species were formerly described from the coast of northern Argentina and are new records for the Magellanic region; one species was formerly known from Antarctic regions.Eight species are still not identified to species level.
The highest species richness can be reported for the family of Nannastacidae (11 species).However, the most abundant species belong to the families of the Diastylidae and the Lampropidae: Leptostylis cf.vemae and Hemilamprops ultimaespei.The highest densities exceeded several thousand specimens per station and were found at stations 1253 and 1213 (Table 2).
The total number of species caught by the epibenthic sledge was similar to the number sampled by dredging (24 and 22) (Table 3).The comparison between the two sampling devices shows a good level of congruence at the species level.Cyclaspis alba was only found in the dredge samples whereas Cumella sp. 2 was only collected in the sledge samples.However, the sledge yielded a much higher number of specimens and species per sample than the dredge.The highest number of species per sample in the sledge samples was 17 (station 1253), in the dredge samples it was 12 (station 1124).The highest number of specimens in the sledge samples by far exceed 1500 (stations 1213 and 1253), whereas the dredge sample maximum was 529 individuals per catch (station 1124, Arntz and Gorny, 1996).
Interestingly, the samples from the Beagle Channel contain two species of the genus Eudorella, which are probably new to science.These two Eudorella species are definitely not identical with those species described from the Antarctic.

Zoogeography
As the Scotia Arc might serve as some sort of a "bridge" between the Magellan area and the Antarctic Peninsula one might expect a clear species correspondence between these regions.
Until now 87 Cumacea species are recorded from the entire area.Only seven of these 87 are found both in the Magellan region and in Antarctic waters.It is evident that the overlap for this crustacean taxon is less than 10 %.While 52 % of the species are endemic for the Magellan region, the endemism percentage of the Antarctic region is even higher (75%) (Table 4 and 5).
A more detailed view of cumacean species from different regions of Antarctica shows some close relationships (species overlap) between subareas (Table 6).For example the Cumacea composition of South Georgia, the South Orkneys, the South Shetland Islands and the Antarctic Peninsula shows a high species overlap (48 to 57%).However, other 298 U. MÜHLENHARDT-SIEGEL

DISCUSSION
Table 5 implies that the cumacean fauna of the Antarctic Peninsula would be more similar to the Eastern Antarctic than to the adjacent Weddell Sea.It seems more likely that this is an artefact, a gap in our knowledge, than a zoogeographical problem.These observed differences in species correspondence might possibly still be due to our lack of knowledge.The genus Eudorella is well represented in boreal areas of the northern hemisphere with 23 species whereas only 4 species are described for Antarctic regions, but not a single 300 U. MÜHLENHARDT-SIEGEL  species of this genus was reported from the Magellan region until now.
A good example is the species Eudorella fallax which was recorded to be endemic to South Georgia (Table 4) in the past.Data from a series of samples from the South Orkney Islands, the South Shetland Islands, the Antarctic Peninsula, and from the Prydz Bay area, East Antarctic (Mühlenhardt-Siegel, unpublished data) show that Eudorella fallax occurs regularly in samples from all areas mentioned.This species can therefore no longer be regarded as endemic to South Georgia, however, it is obviously endemic for the Antarctic.The example of the cumacean genus Eudorella in the Antarctic region has shown the problem of low research effort in parts of the region very clearly.
Maybe a more detailed analysis of other geographical areas with adequate sampling devices like epibenthic or suprabenthic sledges will reduce the number of endemic Cumacea species in the subregions of the Antarctic.The faunal composition of the Cumacea in the waters around the entire Antarctic continent may then show different similarities than those seen at present.
Why is there so little overlap between the Magellan and the Antarctic regions?
The reason for this cumacean zoological feature may be found in the biology of these peracarids: -Cumacea are living preferably in soft sediments or hyperbenthically, only occasionally moving into the water column, -they are brood protecting, the early life stages are kept and sheltered in a marsupium.
Therefore there is almost no opportunity to drift and disperse over long distances and thus deep-sea areas become effective geographical barriers for shelf species.
The species composition of the Cumacea in the Magellan region and in Antarctic waters leads to the conclusion that Antarctica was probably not colonized from South America along the Scotia Arc.However, the origin of Antarctic Cumacea is still not solved.It is possible that the Antarctic Cumacea originate from an ancient Gondwana fauna or they might be closely related phylogenetically to the deep-sea fauna as are the Serolidae or Arcturidae of the Isopoda (Brandt, 1991(Brandt, , 1992)), or the Iphimediidae of the Amphipoda (Watling and Thurston, 1989).However, hardly anything is known about Cumacea from the deep-sea close to and around the Antarctic continent.Therefore more deep sea sampling around Antarctica, from off the continental shelf, downslope towards the abyssal plains is needed in order to solve this question.

TABLE 1 .
Arntz and Gorny (1996)Joint Magellan" Expedition with RV "Victor Hensen" in 1994: epibenthic sledge stations according to depths.The first two figures in the station list are the author's counts, the following four figures refer to the station list presented byArntz and Gorny (1996).

TABLE 4 .
-Cumacea of the Magellan and the Antarctic/Subantarctic regions.(+): also described for north of 40°S; n: new record for this region; A: Argentina; C: Chile; F: Falkland Isl.

TABLE 4 .
(cont.) -Cumacea of the Magellan and the Antarctic/Subantarctic regions.(+): also described for north of 40°S; n: new record for this region; A: Argentina; C: Chile; F: Falkland Isl.