28 G. D. FULLER AND A. L. BAKKE 



. Considering the difficulties mentioned above, before we make 

 use of the variations in leaf size to characterize a given forma- 

 tion we must distinguish between and determine the number of : 



1. Species with simple (including lobed) leaves, 



2, Species with (a) lacerated leaves and (b) dissected leaves. 

 It is apparent that the size of the three groups will give some 



interesting data concerning the formation under consideration. 



It is necessary to treat every life-form by itself; for example, 

 the evergreen cannot be treated in the same group with the 

 deciduous mesophanerophyte, etc. After the material is di- 

 vided into groups, according to their life-forms, the "leaf-size" 

 of each species in all the groups, or in one or more of them, de- 

 pending upon which are the most important, is determined. 



If the arborescent vegetation of a tropical phanerophyte for- 

 mation is examined and found to consist of (1) 1% deciduous 

 phanerophytes, (2) 99% evergreen phanerophytes, in which (a) 

 9% have compound, lacerated and dissected leaf-blades, (b) 

 90% have simple (including lobed) leaves, then the first thing 

 will be to determine the "leaf-sizes" of the species in this last 

 and largest group. 



It will be easier and more significant to investigate the group 

 or groups with the simple (including lobed) leaves for generally 

 these will be far the most numerous in all phanerophyte and 

 chamaephyte formations. Consequently I shall confine atten- 

 tion to this group and determine the "leaf -size" for every 

 species. 



An exact determination of the area of leaf surface cannot be 

 undertaken because the leaves of the same plant will vary more 

 or less in size and also because such measurements would take 

 much longer than the importance of the results would warrant, 

 rather in all practical comparisons of leaf-size it will be neces- 

 sary to separate the leaves into a number of classes according 

 to size; and for all ordinary cases there must not be too many 

 classes. It should be very easy, from the very outset, to refer 

 a plant to the proper class based upon "leaf-size." 



The primary questions are: (1) How many "size-classes?' and 

 (2) What are their limitations? I am in favor of a few classes 



