3-Ketosteroid methoximes (MO derivatives) undergo a degradation when injected into columns that contain acidic materials or supports with acidic sites. Degradation products can usually not be isolated; they result from a Beckmann fission reaction. The extent of sample loss is dependent on the amount of acidic material (pyridine hydrochlo-ride), and decreases with time as the column is “healed” by sweeping or silylation treatment. Lanost-8-ene-3-one MO derivative, where the methoxime group is adjacent to a tertiary carbon atom, gives an unsaturated nitrile by Beckmann fission. An appropriate analytical technique was developed to prevent sample loss. About 1–3 cm of 10% coated support at the head of the column will prevent permanent damage to the column; the solution of the analytical sample should also contain an excess of silylation reagent. When this is done, no sample loss is observed in GC studies of adrenocortical steroid hormones as MO-TMS derivatives.