12 REPORT OF COMMISSION VII 
previously-mentioned extremes. Almost any regional vegetation group consti- 
tutes such a happy medium because (1) the classes to be recognized are many, 
but not innumerable, and (2) the variation within each class is appreciable, 
but not completely random, thanks to the presence of chromosomes in the vege- 
tation which transmit genetic characteristics from generation to generation 
in accordance with Mendel's laws of heredity. 
The group selected for illustration of various kinds of keys, namely ''Wild 
Vegetation of the Tropical Pacific" is presented here with some reluctance be- 
cause it partially duplicates material which has been published in several pre- 
vious articles. However, there are no military security problems in discussing 
this group, the photo recognition features of the group are well known, these 
features have already been organized into several of the previously-mentioned 
types of keys, and they have been tested on large numbers of photo-interpreta- 
tion students. It is believed that in these respects no other group will serve our 
purpose so well. 
One portion of an essay key for this group appears in Figure 5. In Figures 
6, 7, 8, and 9 are presented portions of integrated-selective, disc, punch card, 
and dichotomous keys, respectively for this same group. 
FIG. 5. PORTION OF AN ESSAY TYPE OF PHOTO-INTERPRETATION KEY. 
In classifying wild vegetation of the tropical Pacific area the photo interpreter 
should first ask himself which of the possible classifications that might be made from 
photographs is the most useful for the purpose intended. For example, there would be no 
point in identifying every tree of a rain forest as to its exact species (assuming for the 
moment that this were possible on aerial photographs) unless some use were to be made of 
this detailed information. In some cases, either for civil or military purposes, it is 
both possible and desirable to classify wild vegetation of the tropical Pacific on a spe- 
cies by species basis. For example, if the objective is to provide the explorer or the 
soldier with information as to the ease with which he could traverse various areas, it is 
desirable to map such species as mangrove (indicating a brackish water swamp), sago palm 
(indicating a fresh water swamp), and casuarina (indicating a firm sandy beach). In other 
cases a grouping of the indicator species of vegetation into broad types. within each of 
which conditions are relatively homogeneous will better serve the purpose of the user. 
For example, moss forest, rain forest, swamp forest and savannah forest constitute an ade- 
quate classification of these broad types for the previously stated purpose. Finally, an 
even more arbitrary system of classification may be both necessary and sufficient for other 
types of vegetation. For example, all wild grasses in the tropical Pacific might logically 
be classified either as tall grass or short grass, regardless of species, depending upon 
whether the grass is taller than a man's head thereby restricting both his visibility and 
the ease with which he might traverse the area. 
By applying each of the above considerations to wild vegetation of the tropical Pa- 
cific the following twelve major vegetation types are recognized: 1. Nipa Palm, 2. Sago 
Palm, 3. Mangrove, A. Swamp Forest, 5. Rain Forest, 6. Moss Forest, 7. Savannah Forest, 
8. Tall Grass, 9. Short Grass, 10. Wild Sugar Cane, 11. Casuarina, 12. Secondary Growth. 
Certain of the aforementioned types, namely nipa palm, sago palm, mangrove and wild 
sugar cane characteristically occupy swampy sites and therefore indicate poor traffic- 
ability conditions. Two of these types (nipa and mangrove) ordinarily are confined to 
brackish water areas, while the other two normally are confined to fresh water swamps. 
The nipa palm has frond-like leaves and this feature is quite helpful to the photo interp- 
reter in distinguishing it from mangrove. Although sago palm also has palm frond leaves, 
the fact that its leaves are arranged in distinct rosettes while those of the nipa are not, 
usually permits the interpreter to distinguish between the two, as shown in the accompany- 
ing figure, 
It is felt unnecessary to present here an illustration of either the file key or 
the photo index key, because either might be illustrated simply by means of one 
of the photographs appearing in Figure 6, together with appropriate notes 
concerning its interpretation. 
Based upon actual tests, it can be stated with confidence that for this