118 
the south and west respectively, are different basically because 
they contain mixed coniferous-deciduous forests rather than closed 
canopy black spruce (see Fig. 6). 
- Section B.10, the lake Nipigon basin which is entirely surrounded 
by higher land may well be distinguishable from surrounding 
Sections simply on the basis of physiographic differences. Such 
differences, combined with the rather uniform, predominantly 
black spruce forest, may also contribute to detection and delinea 
tion from surrounding Sections. Section B.10 was also detected 
and delineated on the NOAA-2 image (see Fig. 7). 
The following additional boundaries as determined by Rowe were 
recognized only on the NOAA-2 image: 
- Section B.8, a relatively level plateau with a predominant cover 
of jack pine (.Finns bariksiana Lamb.) with black spruce in mixture 
or predominating on poorly drained soils. 
Separation from Sections B.22a to the northwest and 
B.ll to the southwest is probably due to more rugged relief 
in both, and to the uniform forests of the former (see Fig. 7). 
In an attempt to increase the amount of information provided by the 
maps produced, the forest boundaries as recognized on ultra-small-scale 
photography (Nielsen and Wightman, 1971) were transferred onto the mosaics. 
It was hoped that these boundaries could then be recognized and further 
followed across the mosaic. The attempt failed. 
CONCLUSIONS 
Although ERTS-1 imagery permitted mapping of the boundaries of some 
of the major forest types, its use in this application is more limited than 
had been expected. A significant factor in this limitation was the 
inconsistency of the imagery. Differences in quality were very pronounced 
among black-and-white as well as among colour scenes. These qualitative 
differences were due primarily to a combination of variables related to image 
reproduction and date of data acquisition. 
These problems do not appear when much smaller scale images such as 
those obtained daily by NOAA-2 are used. The homogeneity of the image over 
large areas is a definite asset. 
It is stressed again that the images used for the study were 
standard products as provided to any user in Canada. Furthermore, the 
conventional photo-interpretation method used may have limitations, but at 
this time, appears to be the only realistic approach for such an undertaking. 
As more imagery becomes available and quality controlled mosaics are produced, 
it is conceivable that the results of classifications of this sort will 
improve.