=Estimation of Exposure.=—According to the foregoing argument the problem of estimating an aerial exposure resolves itself largely into one of deciding how short this may be made. Or, if the light is strong, whether it is sufficient so that a filter may be introduced without demanding more than the 1/100 second or thereabouts which is dictated by the motion of the plane. Deciding upon exposures in the field has been largely a matter of experience and judgment. A majority of the cameras in use during the war were not furnished with shutters calibrated in definite speeds. Consequently, the sergeant upon whom the decision usually devolved became a storehouse of knowledge as to the slit widths and tensions appropriate to each individual camera. This knowledge had to be acquired from the results of actual photographic reconnaissances, or from special test flights, both of them wasteful methods. But the chief objection to this state of affairs lies in the fact that the knowledge thus acquired is of no use to anyone else, nor is it applicable to other types of camera. The first essential to placing exposure estimation upon a sound basis is therefore an accurate knowledge of shutter performances. Either the shutter speeds should be placed upon the camera by the manufacturer and periodically checked, or a regular practice should be followed of calibrating shutters, either at a base laboratory or even in the field. Assuming that the speeds of all shutters are accurately known, the process of estimating the requisite exposure becomes less a matter of mere guesswork and more nearly a matter of precision. For this purpose data on the variation of light intensity during the day and during the year (Figs. 101 and 102) should be taken as a guide. These data refer of course to visual and not to photographic light, but since it is always necessary to use color filters, which make the active light of approximately visual quality, this is no valid objection. The effects of clouds and mist must of course be learned largely by experience, but with the above daylight data at hand, anyone in possession of definite information on the correct exposure with a given plate for a known day and hour need not go far wrong in estimating exposures at any other time in definite fractions of a second. [Illustration: FIG. 109.—Chart showing aerial exposures for all times of the day and year. Data on basis of F/5.6 lens, Jougla orthochromatic plate, and clear sunlight, no filter. Exposures to be doubled and tripled for overcast and cloudy weather.]