252 REPORT OF THE No. 3 



pictures are valuable as permanent records of important aerial observations 

 and as a substitute for lengthy descriptions of forest conditions. For mapping 

 purposes, however, they are of little use. 



Vertical pictures, on the other hand, while they are not artistic or interest- 

 ing at first sight as pictures, can be used for mapping and type classifying. 

 The cameras making such exposures are, however, delicate and complicated 

 instruments. In operation, they are fixed to the machine, and can be made 

 to produce, either automatically or by a simple operation, exposures at such a 

 rate that each succeeding exposure will cover a certain proportion (greater or 

 less as desired) of the territory covered by the exposure preceding it. By thus 

 including a part of one picture in another — or "overlapping" — it becomes 

 possible to "join" one to another and build up a composite picture containing 

 any number of single exposures. The accurate joining of two pictures of this 

 kind is not, however, as simple a matter as might be supposed from the fore- 

 going, which takes no account of the motion of the machine during exposure. 

 Such motion, the effects of which must be corrected by dark room work and 

 which might be compared to that of a ship — sometimes slowly rising and falling, 

 as if riding on the long, gentle swell of an aerial ocean ; sometimes rolled sharply 

 from side to side or pitched fore and aft — continually and unavoidably violates 

 the requirements for exact similarity in the scale of overlapping pictures — 

 namely, that exposure be made at exactly the same height and with fore and 

 aft and lateral axis either level or at the same degree of departure from level. 

 While the differences arising from the above causes are usually small in each 

 separate case, the total efi^ect over a series of hundreds of pictures may pro- 

 duce great inaccuracy. Nor is the motion of the machine the only element 

 affecting the accuracy of aerial photographic survey, for even in perfectly built 

 up strips, where point has been made to correspond with point exactly in the 

 overlap, there is still the error due to the natural differences in ground eleva- 

 tion. Unfortunately, such errors require for their correction, besides labor- 

 atory work, a certain amount of ground survey work, which shall be visible 

 in the picture and whose details can be plotted and used as a check. Errors 

 arising from the foregoing however, do not have to be seriously reckoned with 

 in a flat country, and where absolute accuracy is not essential. 



Results of this year's work would indicate that satisfactory pictures can 

 be produced which will show sufficient detail to separate types; but it would 

 appear that before this possibility can be utilized to any great extent, it will 

 be necessary to accumulate a photographic reference library of known forest 

 growths as a basis of comparison, and also to devise a satisfactory method of 

 joining the strips of pictures to one another to make a mosaic. Such joining 

 is not a matter of the operation of a mechanical device, as in the joining of 

 successive pictures to make a strip, but is dependent on the ability of the pilot 

 to fly his machine along evenly spaced parallel courses, so as to cause the long 

 edges of the strips to join — a very different and more difficult undertaking. 

 Corrections for differences in ground elevation and motion of the machine 

 must be made in this connection, just as with the sections of each strip. 



Despite the difficulties outlined above, the use of aerial photography for 

 mapping would appear to be an almost certain development in survey work. 

 Its ability to reproduce detail is unlimited: it can also produce results rapidly 

 and with a degree of completeness, the value of which can hardly be appreciated 

 without experience. 



8. Fire -Pa/ro/.— Detection of forest fires was carried on during the fire 



