23 



respect of height and shape, i.e., the case of greater or less diffi- 

 culty in selecting correct representatives of the class in question, 

 and (iii), on the possihility of errore due to an insaffieient number 

 selected in one or more classes being compensated by excess or 

 deficiency of restdts obtained for the other classes. Obviously the 

 possibility of such compensation exists only if each sample stem 

 represents approximately an equal fraction of the total contents of 

 the crop. 



As a rule, the number of sample trees measured in any girth- 

 elass is determined by the proportion either (a) of the nnmher 

 of stems, or (4) of the aggregate basal sectional areas comprised in 

 1 hat class. The adoption of the latter of these two bases of calcu- 

 lation is always more likely to give correct results. 



The grouping of the trees into height-classes must take place 

 at the time of the enumeration survey, and be a part of that 

 operation ; their separation into girth-classes may be effected after 

 the survey with the help of the figures which it furnishes. 



To illustrate what has been said in this section we append 

 l>elow a specimen of a survey by girth-classes. 



Forest Eanjatra Sample Plot (a) 264' x 165'. Age 135 years. 

 Compartment 97. Area 1 acre. Density of stock 0"9. 



