Out of 100 microliters of capillary blood of 32 normotonic and 55 hypertonic (RR at least 140/100) diabetic patients, the following parameters were determined at 09:00 a.m.: pH, pCOz, baseexcess, HCO3, PO2,02sat., Na, Ca, Mg, K, lactate and blood glucose. Group averages of the normotonic and hypertonic patients did not show any significant differences in the parameters mentioned. But when we split the 2 groups along their glueose averages, forming 4 subgroups with higher and lower glueose levels, significant differences between Mg and K in the respective groups were evident, pointing towards an increase in electrolyte metabolism due to both hyperglyeemia and hypertonia. Linear regressions between Mg, glueose and pH within the four groups inereased eircumstantial evidenee for our supposition of increased electrolyte metabolism being attributable to both hyperglyeemia and hypertonia, considerably. It turned out that high blood glucose alone already increases electrolyte metabolism, shown by positive correlations with Mg and glucose and negative correlations between Mg and pH which culminates in the highest Mg values along with highest glueose and lowest pH values in patients with eombined hyperglyeemia and hypertonia. Although we partly drew similar conclusions in a fonner paper [14], the step forward in the present work is that results won by eomparison of parameter changing (delta values) between two sampling events can be confirmed by using the directly determined data of just a single sample. Thus, appropriately processed data and eomparatively simple mathematical operations like implementation of linear regressions do not only detect significant differenees hidden behind similar group averages, but also con-siderably curtail the necessary diagnostie blood sampling.