This paper examines the assumption that the audible events detected as Korotkov sounds in sphygmomanometry occur when blood pressure equals arm-cuff pressure. Several effects that contribute to discrepancy between these pressures are quantified using an idealised arm-and-cuff system consisting of a thick-walled collapsible tube subject to external compression along a central part of its length. The effects studied are (1) transverse pressure difference, resulting from tissues sustaining a part of the external compression through (a) circumferential bending stiffness and (b) longitudinal curvature of the tensed localised neck at the site of initial collapse, (2) longitudinal pressure difference between upstream pressure and pressure at the collapse point due to both (a) viscous and (b) inertialpressure drop. These effects are found to compensate partially for each other; the pressure within the vessel at the collapse point is less than the cuff pressure, but is also less than the blood pressure at the upstream end of the cuff. All four of the contributing terms increase proportionally to the flow-rate raised to a power greater than one, except the viscous pressure drop. Owing to a progressive shortening of the collapsed neck as flo w-rate increases, the viscous term is almost independent of the flow-rate. The overall discrepancy displays less flow-rate dependency and is smaller than some of the terms which contribute to it. This means that considerable accuracy is needed if measurements of the effects are to be used to correct the raw data on cuff pressure at the time of Korotkov sound emission so as to obtain an improved estimate of the blood pressure.