Summary Photogrammetry as applied to the establishment of colonial maps on medium scales (1 : 25.000 – 1 : 50.000) includes aerial triangulation and aerial levelling. Planimetric ground control can be provided by geodetic triangulation or by a limited number of oriented baselines, the longitudinal deformations of the strips being regular. Altimetric ground control on the other hand must be relatively dense especially in view of the transversal torsion of the strips, which is generally irregular. Therefore it is necessary to determine the heights of a relatively large number of ground control points situated outside the axis of each strip. These heights can be determined easily by barometric levelling. A new accurate aneroid barometer, designed by the author is described. It consists of a column of seven metal boxes C 1, C 2 … (fig. 1, page 25 fig. 2, page 26) whose deformations add together. The movements of the surface of the upper box, due to changes of atmospheric pressure are not — as in ordinary aneroid barometers — magnified by a system of levers, but observed directly by optical means. A wire F 1(fig. 2.) whose movements are identical with those of the surface of the upper box, is observed by a 14 power microscope M.C. (fig. 2) or H 1 H 2 (fig. 1). A fixed wire F 2 is observed simultaneously. The variation of the distance between these wires is measured by means of an eyepiece micrometer M, estimating 0,001 mm. A thermometer permits the measurement of the temeprature of the compartment which contains the metal boxes. The whole is mounted on rubber cushions in a wooden case. The case is suspended fibration free from a metallic frame by means of rubber tubes. (fig. 3). A description is given of trials carried out in Belgium. The instrumental constant was determined each day by measuring three datum stations. The atmospheric pressure at these points was measured during the day every ten minutes by means of mercury barometers and reduced to 20° temperature. (fig. 4, 5 and 6, pages 29–30). The tables T 1, T 2 and T 3 (pages 32–34) fourth column show the altitudes measured with the new barometer at a number of points. The fifth column gives the differences between these results and the geodetic altitudes. These differences are affected by a systematic error, caused by the general change in atmospheric pressure in the area concerned. This change is represented by straight lines in fig. 4, 5 and 6, indicating sections at time AB, BC and CD. The systematic error mentioned was eliminated by computing the mean of the differences per section and subtracting it from each difference. The adjusted differences thus obtained are shown in the seventh column of the tables T 1, T 2 and T 3. Their mean value E (written at the bottom) is a function of e 1, e 2 and e 3: E = e 1 2 + e 2 2 + e 3 2 in which: e 1 is the mean error of the pressure measured at a datum station e 2 is the mean error of the pressure measured at the other points e 3 is the mean error caused by slope of the pressure level. e 1 was computed by comparing the measurements at the three datum stations, to give e 1 = 0,4 m. The error e 3 is assumed to be a linear function of the fluctuation of the pressure observed at the datum stations, with respect to the straight lines AB, BC and CD (fig. 4, 5 and 6). These fluctuations are: June 3: 0,8 m whence e 1 = 0,8 k June 5: 0,7 m whence e 1 = 0,7 k June 7: 1,1 m whence e 1 = 1,1 k Substitution of these values in the above equation gives three equations, the solution of which by the method of least squares gives the internal mean error of the instrument: e 3 = 0,4 m In equatorial regions the mean error e 3 would be negligible. Assuming a datum station provided with two barometers of the type described and four carried to the other points, the mean error of the altitude may be estimated at 0,2 m.