. 24 



The preceding detailed study indicates the several cases in which 

 it is necessary to form only height-classes or ouly girth -classes, 

 or comhiued height and girth-classes. Every crop to be sur- 

 veyed must be divided up in this manner into a small number of 

 distinct groups, for each of which the corresponding sample tree, 

 possessing the average basal section, may be taken as a true repre- 

 sentative in respect of both height and form-factor. 



In the constitution of girth-classes one of three systems may be 

 adopted, riz. 



I. THAT OF EQUAL GIRTH-GRADATIONS. The number of grada- 

 tions are usually from three to five, and the figures of the 

 enumeration survey at once indicate the number of trees 

 to be included in each gradation. In this system the 

 disparity between the numbers of trees in the several 

 classes is always very great. 



II. THAT OF NATURAL SIZE-CLASSES. In the most canopied 

 crops at least three classes of individuals are easily recog- 

 nisable (1), a class of small stems with restricted crowns 

 ana below the average height of the crop; (2), a middle 

 class with freely developed crowns and of the same height 

 as the crop ; and (3), a class of thicker stems with crowns 

 rising above the general mass of the crop. To these classes 

 may be added two others, viz., that of suppressed indi- 

 viduals and that of old standards or trees of much o-reater 

 age than the rest of the crop, arvd conspicuous by their re- 

 latively great size. These several classes are generally at 

 once recognisable from the figures of the enumeration 

 survey. 



Ill, THAT OF EQUAL GROUPS. If ,natifral size-classes are not 

 distinguishable by their relative thickness of stem, then 

 it ie advisable to form classes containing approximately 

 the same number of stems. In 4his method the range of 

 girth included will obviously differ from class to class, 

 and the method is therefore to this extent the very op- 

 posite of I. 



The numher of sample trees to measure in each girth-class 

 will depend (i), on the relative proportion to the entire crop which 

 the claes represents either in respect of number of stems or of cubi- 

 cul contents; (ii), on the uniformity of the component trees in 



