26 
BSH Report No. 56 
Considering the results of the duplicate analysis, the data consistency within 
the laboratories measurements appears therefore to be good. Comparing both 
laboratories, using absolute values however, reveals a rather different situation: 
Looking at both 13 C-content averages we observed an absolute discrepancy of 
3%o. Consider discussion point DP-2 on page 31. 
In the context of paraffin-wax samples one could conclude that this discrep 
ancy of actually Identical samples originate from different products which are 
derived from different raw materials (oil)/or different sources. Thus, in case of 
spill Identification, this could be contradicting the otherwise solid results from 
GC-FID and GC-MS analysis. We therefore conclude that it might be possible 
that samples which appear to be similar In terms of GC-FID results (bell 
shape), nonetheless might be derived from different materials and conse 
quently do not stem from the same source. 
At this point it becomes obvious that there is some need to compare methodi 
cal and Instrumental properties in order to Increase inter-laboratory accuracy. 
As shown In Tab. 5 is a high discrimination power between the samples using 
IRMS. From the matching ratios, one can derive a high probability for samples 
of set 1 to originate from the same bulk or educt material. This is of course a 
necessity for having samples of the same composition. However, It is also pos 
sible that the samples consist of two different products or batches, which were 
produced from the same origin material. 
To sum up, this technique is undeniably a valuable addition to spill Identifica 
tion Investigations, but it has to be interpreted in context of results from e. g. 
GC-FID. 
Consider discussion points DP-5 and DP-6 on page 31.