It is known that the positive column is unstable in a strong longitudinal magnetic field. The correct explanation of this phenomenon has been given by Kadomtsev and Nedospasov. However, the instability mechanism is not fully interpreted in their work. In this contribution, the main features of Kadomtsev and Nedospasov's stability criterion are derived from simple physical arguments, and the mechanism responsible for instability is thus shown in a quantitative way. In particular, the perturbed radial potential distribution can be related to the perturbed density distribution, and hence need not be assumed. It turns out that this screw instability is a new and important type of instability hitherto not considered. The instability mechanism is not very sensitive to the degree of ionization, and hence may also be expected to be operative in highly ionized plasma columns. Possible connections with the instabilities observed in the Stellarator and the Zeta machine are discussed. An attempt is also made to explain regular oscillations observed in a toroidal discharge in terms of the present theory.