results from original diapositive be accepted before results from 
duplicate diapositives, and results from paper prints should be accepted 
as the least reliable of the three choices. In addition for purposes of 
vegetation damage inventory and monitoring, results from color-infrared 
photos should be accepted before results from normal color photos. 
There are few, if any reports in the literature to support a case for 
total dependency of remote sensing for vegetation damage detection on 
normal-color film alone. However, there are the odd exceptions where 
a special technique must be used (Bradshaw 1974). 
E. ‚Evaluation 
One final problem needs to be mentioned. In the evaluation of 
damage, there is a need to develop a technique for air photo determination 
of the actual crop values lost in monetary terms. Some work has been 
done in Agriculture (Wallen and Jackson 1971) and tree volume table 
adjustments have been suggested in forestry (Weber 1965), but nothing 
has ever really taken hold and become an accepted practice for damage 
evaluation. In forestry studies, the numbers of dead trees are counted 
or estimated on the air photos, then a ground sample is taken to deter- 
mine volumes affected. Frequently the volume estimation is restricted 
to the dead tree volume. Is it possible to determine volume loss due 
to partial defoliation? It is recommended that controlled, analytical 
studies be done to determine the relation between measurable air photo 
variables such as crown density, crown area, etc., and related to vege- 
tation parameters, as they can be measured on the ground. Much work is 
needed in this aspect of remote sensing for vegetation damage. 
SUMMARY AND RECOMMENDATIONS 
Many remote sensing and photo interpretation techniques have 
been suggested and successfully tested for the inventory or monitoring 
of damages to vegetation. Problems have arisen partially because of 
differences in definition of terms, and a lack of an integrated philos- 
ophy with respect to vegetation damage and remote sensing. 
1. It is recommended that a universal method of classifying damage to 
vegetation be adopted to avoid the confusion of many descriptive terms, 
and to provide a means for comparison of results. Because a given 
damage type may be caused by any one of a number of agents, or conversely, 
any one agent can cause many types of damage, there is a need to employ 
the holistic approach to initial damage detection. 
II.It is recommended that a library of information on damage types that 
any given agent may cause be collected. Particular attention should be 
paid to spectral reflectance signatures for possible future application 
to multispectral analysis.