Chlordane components (CHLs) and their metabolites(heptachlor,
cis-heptachlor epoxide, U82, MC4,
trans-chlordane, MC5,
cis-chlordane, MC7, oxychlordane, MC6,and
trans- and
cis-nonachlor) and aldrin, dieldrin, endrin,isodrin, endosulfan 1, endosulfan 2, and mirex were quantifiedin the soft tissues of blue mussel, a whole crab, andwhole fishes collected from the spatially different sites inthe Gulf of Gda
sk. Six to twelve chlordane compoundsand metabolites and dieldrin were detected in all organismsexamined while aldrin, endrin, isodrin, endosulfans 1 and2, and mirex were not found above the detection limit of themethod. The lipid weight based concentrations in Balticbiota were relatively small and ranged from 12 to 150 and 7.6-77 ng/g, while between 0.16 and 6.8 and 0.10-6.6 ng/g infresh tissue, respectively. The profile (%) of chlordanecompounds was very similar between various fish specieswith
trans-nonachlor (28 ± 17),
cis-chlordane (23 ± 18),oxychlordane (13 ± 7), and heptachlor epoxide (11 ± 5) asmajor constituents and was totally different in crab withoxychlordne as the most dominating (>65%) compound. Bluemussel, lamprey, and three-spined stickleback exhibiteda smallest ability to metabolize CHLs, and such fishes as cod,lesser sand-eel, sand-eel, pikeperch, perch, round goby,flounder, and herring showed a slightly better ability, whilecrab was able to effectively metabolize most of CHLcompounds except
trans-nonachlor. A value of the quotientof the
trans-nonachlor to
cis-chlordane concentrations(N/C quotient) was 1.0 in blue mussel, 3.1 in crab, andbetween 0.9 and 1.8 in fish. Both the small concentrationsof CHLs in all organisms and the values of N/C quotientsclose to 1 imply on a long-range aerial transport throughmovement of the air masses from the remote regions of thenorthern hemisphere as a main source of this pesticidein the Gulf of Gda
sk. The interdependences between theCHL profiles for various fish species and betweendifferent sampling sites were examined using the principalcomponent analysis (PCA) method. Applying the PCAmodel the first four significant components explained 90%(43% + 23% + 15% + 8%) of the total variance in thedata matrix.