An analysis was carried out on 51 samples from 32 stations, obtained during
the I.C.O.T.A, cruise between 18 January 1996 and 7 February 1996 in the P
ointe Geologie Archipelago, Terre Adelie, and on samples obtained by 2 sedi
ment traps deployed in the Ross Sea, within the framework of the Italian C.
L.I.M.A. Project. These samples showed the unexpected presence of numerous
larvae of benthic marine invertebrates. In addition, larvae from the Ross S
ea were examined. Seven phyla with complex life-cycles and at least 18 spec
ies of benthic polychaetes were represented. Morphological descriptions by
scanning electron microscopy of the most frequent larvae and identification
of larval types are provided. Descriptions from the literature of larvae c
ollected in the Antarctic Peninsula were used for comparison. Morphological
descriptions do not always lead to identification at the species level; ho
wever, these descriptions are indispensable for establishing the geographic
distribution of larvae. It has been demonstrated that indirect identificat
ion methods relying on the premise that the benthic fauna is entirely known
and that sampled larvae belong only to species already known through their
benthic adult stage may lead to obvious absurdities. As an ecological cons
equence, Thorson's rule, according to which indirect development should not
be represented and direct development should be dominant at high latitudes
, is not supported by our data. A short comparison of the local reproductiv
e pattern with a subantarctic area, the Kerguelen Archipelago, strongly sug
gests that the absence of planktonic development in this area cannot be exp
lained by Thorson's rule. This rule has been invalidated in more austral zo
nes: the prevalent reproductive pattern in the Kerguelen Archipelago is lin
ked to the very marked advective strength of water flow.