Filling compound effects, analysis component

We computed an analysis of variance over Table 1, followed by tests of specific effects. Two results were statistically significant. Random selection of compounds was better than cluster or space-filling selection. BCUT descriptors were better for analysis than either of the principal component descriptor sets. 

Filling the insert with silanized glass wool has proved effective for the absorption of the sample hquid and for rapid heat exchange. However, the glass wool clearly contributes to enlarging the active surface area of the injector and should therefore be used only in the analysis of noncritical compounds (alkanes, chlorinated hydrocarbons, etc.). For the injection of polar or basic components special care is required keeping the glass wool deactivated by frequent replacements, an empty deactivated insert is recommended. 

---