Annual mean values of the north (X), east (Y) and vertical (Z) components of the magnetic field from 20 observatories in Canada and adjacent countries for the period 1955–1976 are used to test secular change derived from cubic time terms in a polynomial model. This model was used to derive the main magnetic field charts of Canada for 1975. The rms deviation in the secular change for 1975 (annual change) is a disappointing 16 nT. Graphs of observed and computed annual means show that this fit will worsen with time, especially in Y and Z. For these reasons the annual change contours on the above-mentioned charts were derived by conventional graphical methods. Graphs of the magnetic declination (D), the horizontal intensity (H) and Z were drawn for 20 observatories and 157 Canadian and American repeat stations. Values were extrapolated to 1975 and contoured. A comparison of graphical and polynomial methods show that this simple graphical extrapolation is more sensitive to small significant secular changes than a polynomial extrapolation. We show annual change charts in X, Y and Z constructed by this method.The movement of centres of intense secular change, called isoporic foci, in X, Y and Z, are examined to determine their influence on the secular change pattern in Canada. These indicate that Y and D will continue to decrease, while X and H will continue to increase over most of the country. The total intensity (F), the magnetic inclination (I), and Z will all decrease rapidly in southern Canada.