The results obtained from a neutron-moderated detector aboard a commercial airline flight between Johannesburg and New York in 1976 revealed a mismatch between the counting rates in the two hemispheres when they were plotted against vertical cutoff rigidity values interpolated from the 1975 values of the 5¿¿15¿ world grid. Cutoff rigidities in the North Atlantic had to be increased by approximately 0.5 GV for a good fit. Detailed vertical cutoff rigidity calculations by the trajectory-tracing method along the flight path in the North Atlantic revealed deviations of up to 0.3 GV at ~8 GV between these values and those calculated by interpolation of the 5¿¿15¿ grid. These values resulted in a better, though by no means satisfactory, fit for the data in the two hemispheres. In order to investigate the extent of these deviations, vertical cutoff rigidities were calculated by trajectory tracing for a 1¿¿1¿ grid in the area between 20¿ and 36¿N and 308¿ and 325¿E. The isorigidity contours obtained for this area indicated that the size of the deviations could be >0.25 GV for a longitudinal extent of perhaps more than 12¿. Furthermore, these deviations were found to extend for more than 1 GV in directions perpendicular to the isorigidity contours obtained by interpolation. This result is an indication that a 5¿¿15¿ grid of calculated cutoff rigidities is not sensitive enough, at least for the North Atlantic. When the flight data were plotted against vertical cutoff rigidities calculated for 1980, a better fit was found between data for the two hemispheres at mid-latitudes, thus casting doubt on the 1975 extrapolation from the 1965 international geomagnetic reference field model.