averaged 5 to 15 grams dry weight and were 

 sized using standard sieving and pipette proce- 

 dures (Folk, 1965). Grain size analyses on the 

 CGC Northwind cores have been conducted by 

 the U.S. Naval Oceanographic Office (Andrew 

 and Kravitz, 1968) . 



Because 71 of 88 samples {Edi^- 

 to-Eastivind) analyzed for grain size distribu- 

 tion contained more than 90 percent clay-and 

 silt-sized material with the larger proportion 

 being clay-sized, X-ray examination of the 

 clay-sized (<4 microns) was more appro- 

 priate than optical studies of the coarse 

 (sand-sized) fractions (>62 microns). In ad- 

 dition, preliminary analyses revealed that 

 heavy minerals (S.G. greater than 2.95) made 

 up only about 1 percent of the 62-250 micron 

 fraction in eight surface samples from the 

 East Novaya Zemlya Trough. Thus coarse 

 fraction mineralogy of sediments from the 

 deeper troughs may not be truly representative 

 of the total sediment mineralogy. 



Samples from both brown and gray-green 

 layers were selected from cores E-21 and E-26 

 for the clay mineralogy studies. The clay-sized 

 fraction was separated by settling and oriented 

 slides were prepared by air-drying aliquots on 

 glass slides. The error in quantitative determi- 

 nations introduced by this mounting technique 

 (Gibbs, 1965) renders the results reported 

 here at best semiquantitative. The slides were 

 X-rayed using Ni-filtered Cuka radiation from 

 2° to 30° 28. Each slide was X-rayed again 

 after treatment with ethylene glycol to reveal 

 expandable components. No attempt was made 

 to distinguish chlorite from koalinite al- 

 though both were present in all samples as evi- 

 denced by a broad, often double peak at 3.53 to 

 3.57 A. Peak areas for the basal (001) reflec- 

 tions were measured and used to make semi- 

 quantitative estimates of the proportions of 

 montmorillonite, chlorite/kaolinite and illite in 

 each sample. 



Selected fragments from several of the 

 coarse fractions (>62 microns) were also pul- 

 verized and X-rayed using a Debye-Scherrer 

 powder camera. Zn-filtered Moka radiation was 

 used in some instances to reduce fluorescence 

 of ferromanganese compounds in the samples. 



The 67 surface and 53 subsurface samples 

 selected for geochemical analyses were air-dried 

 and ground in a porcelain mortar to pass an 80- 



mesh sieve. Of each sample, 100 mg. were then 

 treated for about 4 hours with 10 ml. of a solu- 

 tion equivalent to IM-hydroxylamine hydroch- 

 loride and 25 percent (V/V) acetic acid. The 

 details and an evaluation of this procedure are 

 given by Chester and Hughes (1967). This 

 treatment dissolves most of the ferroman- 

 ganese minerals, the carbonate minerals and 

 extracts adsorbed trace metals, but does not af- 

 fect detrital minerals (Chester and Hughes, 

 1967), other than detrital carbonates. After di- 

 lution of the solutions resulting from the above 

 treatment, suitable aliquots were analyzed for 

 iron and manganese using atomic absorption 

 (Beckman Model 1301 coupled to a DB-G 

 Spectrophotometer). Standard solutions were 

 prepared in a dilute HCl matrix and sample 

 matrix efi'ects checked by standard additions. 

 Duplicate and triplicate analyses of selected 

 samples established that the analytical preci- 

 sion was always within ±10 percent and fre- 

 quently within ± 5 percent. 



Several samples were also analyzed for total 

 iron and manganese. After fusing with LiBOa 

 (Shapiro, 1967) the whole sediment was dis- 

 solved in weak HCl and analyzed for iron and 

 manganese using the same standards employed 

 for the nondetrital extractions. Matrix eff'ects 

 were present in the total iron and manganese 

 analyses and the resulting values could only be 

 considered semiquantitative. The concentra- 

 tions obtained for extractable (nondetrital) 

 and total manganese were similar, indicating 

 that manganese is almost entirely associated 

 with the nondetrital fraction. The total iron 

 content, however, was three to four times 

 higher than that found in the nondetrital ex- 

 traction. This is in agreement with the findings 

 of Chester and Hughes (1966, 1967, 1969) and 

 suggests that iron is largely in the silicate or 

 resistant oxide phases. 



Statistical 



The data of the sedimentary properties such 

 as percent gravel, sand, silt, clay and mean 

 grain size, standard deviation, skewness and 

 kurtosis were computed by digital computer 

 based on the results of the grain size analyses. 

 The geochemical data for the surface samples 

 were submitted to a series of linear multiple 

 regression analyses using the computer. Miller 

 (1956) and Krumbein (1959) have discussed 



