APPLICATION OF PHYTOSOCIOLOGICAL TECHNIQUES 293 



Raunkiaer's leaf-size classes have not been widely used, although they are 

 equally amenable to statistical treatment as are the life-form classes. Table 

 15 shows both types of classification simultaneously, not only for the total 

 flora of the sample plot, but also for the sub-classes. It is seen that the 

 mesophyll leaf size is predominant, with 68.35 per cent of the species having 

 leaves of that size range. Only about 11 per cent have larger leaves, while 

 something over 20 per cent have smaller leaves. The leaf-size classes smaller 

 than mesophyll are better represented in the higher strata of the forest than 

 they are in the lower, as shown by the tallest trees (megaphanerophytes) 

 having 4.08 per cent leptophyll, 2.04 per cent nanophyll, and 16.33 per cent 

 microphyll leaves, whereas the medium-sized trees (mesophanerophytes) 

 have 2.:^^, zero, and 10.46 per cent, respectively. There is further discussion 

 of life-form and leaf-size classification in Part III of this paper. 



Discussion. Black et al. (1950) studied three 1-hectare plots of equatorial 

 rain forest, two of which were in Para near our study. On these plots they 

 found 60 species among 564 trees of the varzea type near Belem, 87 species 

 among 423 trees of the terra firme type, and 79 species among 230 trees of 

 the terra firme type near Tefe Amazonas. From a study of their data they 

 concluded that "only about half, or less than half, of the species of trees 

 which exist in a given forest type in the region studied have been met with 

 and recorded on our plots." In a later paper Pires et al. (1953) studied a 

 3.5-hectare plot at Castanhal, about 120 km. east of Belem. Here they found 

 179 species among 1,482 trees 10 cm. or more in diameter. Using the method 

 of Preston (1948), they concluded that more than half the tree species of 

 the local type had seen sampled and that all together the number was prob- 

 ably about 250 species. 



In the present study at Mucambo, which is near the Belem plot mentioned 

 above, we found 153 species of trees 10 cm. or more in diameter, among 897 

 trees. ^ 



Figure 1 gives the species-area curves for the three terra firme plots from 

 Para. Curve A is for the 1950 study of Black et al, curve B for the Castanhal 

 study, and curve C for our Mucambo analysis. In each case the plot was 

 subdivided into 10 X 100 m. strips. The data for these curves are obtained 

 by determining the "new species" added to a list as an additional strip is 

 examined. All three studies can be compared at the 10,000 sq. m. area, which 

 was the maximum size of the early Belem plot. As tested in this way, the 

 floras differ in richness of tree species: 87, 108, and 144 species for the A, B, 

 and C curves, respectively. Castanhal and Mucambo can be tested at the 

 20,000 sq. m. area where the difference is 149 and 173 species for the B and 



■' The original list of Dr. Pires contained some trees later dead and a number that 

 did not quite measure 10 cm. diameter at 4.5 ft. above the ground by our steel 

 tape. The earlier list had 173 species among 1188 trees. 



