APPLICATION OF PHYTOSOCIOLOGICAL TECHNIQUES 325 



evergreen aboveground parts. They often have stolons, rhizomes, or rootstocks 

 and buds that would cause the plants to be classified as cryptophytes or hemi- 

 cryptophytes if it weren't for the persistence of vegetative activity by above- 

 ground parts. In temperate regions such plants die back to the ground, or 

 under it, but in tropical vegetation it seems more accurate to classify such 

 plants as chamaephytes or nanophanerophytes, even though not woody, for 

 they do endure the environmental conditions their aerial parts are exposed 

 to, and that is the crux of the Raunkiaerian system of life-form classification. 



In the classification of plants according to leaf-size classes we have in all 

 cases used the leaflet as the unit in the case of pinnate and palmate compound 

 leaves. When leaves are deeply lobed some judgment is necessary as to 

 whether the lobe is the physiological unit comparable to the blade of the 

 simple leaf. In other words, we have not considered a leaf as simple, though 

 in a strict morphological sense it is, if the divisions are so deep that the lobes 

 have comparable exposure to leaflets. As explained earlier, leaf areas were 

 determined as two-thirds of the length-breadth rectangle as a system of 

 measurement sufficiently accurate in nearly all cases for classification into 

 the Raunkiaerian units. At least any errors of classification made in individual 

 cases, when areas are near the division line between classes, probably counter- 

 balance one another. In all cases measurement is made of the blade, omitting 

 the petiole, but attenuate leaf tips are included in the length measurement. 

 Leaves of a species vary greatly in size in some cases, as between sun and 

 shade forms. We have consistently tried to select representative leaves of the 

 exposed type as the usually larger leaves of protected parts of the plants do 

 not represent the correct relation between leaf size and general climate, but 

 rather the more favorable microclimate of forest interior. This point applies, 

 of course, only to transgressive plants that ultimately attain the canopy or 

 an emergent position. 



The absence of technical determination of the species does not affect the 

 validity of the leaf-size and life-form classification of the taxa. We have, 

 however, done our best to attach correct Latin names to the entities, although 

 the Brazilian flora is extremely rich and imperfectly known. Most of the de- 

 terminations have been made by comparison with herbarium materials in 

 various Brazilian institutions, often with the assistance of local botanists, as 

 indicated in the text. We do not know, however, that the Latin name used 

 is the latest or most accurate. It is for that reason that we have added the 

 name authorities in most cases. Family names are also added as a matter of 

 convenience. There are several cases where further study and recourse to 

 specialists would provide a technical name, but we have not thought that 

 the nature of our study warrants the time and effort involved. 



Table 25 summarizes the Brazilian rain- forest results as to life-form class 

 according to Raunkiaer's five principal classes. In comparison with Raun- 

 kiaer's normal spectrum (which does not necessarily represent the world 



