For the measurements a Hitachi KP 161 CCD camera, a PC with built-in frame grabber and the software package IPPLUS, version 4 were used. Before each measurement series the system has to be calibrated.

The sediment sample was homogenised, sieved over a 2000µm sieve and split. Next, grains from the resulting sample were loosely scattered on a petri dish in such a way that grains are isolated as much as possible. With camera and frame grabber a picture of about 50 grains was produced on the monitor. Next there is a visual screening for grains that are lying against or one top of each other and that could be interpreted by the software as one grain. If so, these grains were separated. 

The picture of about 50 grains was analysed by IPPLUS in such a way that marginal grains were not considered. In this way 10 pictures were studied. The resulting 500 values for the grain surface were put in a spread-sheet. From the surfaces, the volumes of the grains were calculated following Komar & Cui (diameter = 2.sqrt(surface/p); volume = diameter³ ). These volumes are used for the cumulative values. The volume percentages below a certain diameter were calculated by interpolation, for which a straight line in between two relevant points was used as a basis. This implies that when grain size differences in between two consecutive points are larger, errors will also be larger. In practice this is an issue by the coarser phi-values.
 
 

	MPS	Sieve	IPPLUS
72GS14-1  D20	261	207
D50	357	319
D80	509	442
72GS14-2  D20	274	236
D50	376	332
D80	534	449
72GS14-3  D20	242	199	330
D50	332	313	446
D80	468	434	728
72GS14-4  D20	393	307	451
D50	553	469	665
D80	777	768	950
72GS14-5 D20	365	294
D50	517	430
D80	739	721
00DW25-4 D20	238	190	298
D50	387	322	526

Table 1. D20,  D50 and D80 values of grain size distributions using laser counter (MPS), sieve and grain shape (IPPLUS) methods. Samples originate from counterflush borings (GS) and  vibrocores (DW). 

Results

The results obtained from this short study are twofold: a comparison between sampling techniques i.e. between coring and counterflush drilling and the comparison between various analytical methods.

As regards the sampling techniques: the expectation was that the grain size results based on counterflush samples would show coarser values compared to the core samples because of the spill-over of fines out of the sampling basins. Indeed, from the 2 one-metre intervals studied, both intervals showed coarser values for the counterflush sample. Besides, the core samples do show a somewhat higher (although still low) percentage mud than the counterflush samples. Seabed dynamics with time might have caused an appreciable grain size variability.

A comparison of the three analytical methods showed that the laser counter results were systematically coarser than the sieving results i.e. with a factor of 1.13. In earlier, more precise, comparisons using a set of sieves a half-phi