380 JOURNAL OF FORESTRY 



evidence that they have been in the upper crown cover from the start. 

 This is Hkely to be easy for intolerant species, and intolerants will 

 therefore be the best indicators. Even tolerant trees, however, can 

 probably be safely used if carefully chosen, and facility will undoubt- 

 edly come with practice. Beyond the century mark, and considerably 

 in advance of it for many species, especially on the poorer sites, height 

 growth is so slow that it may be regarded as practically at a standstill. 

 The average height of the dominant crown cover of old stands, even 

 when of tolerant species, will therefore probably constitute a good 

 index of the site — the site, of course, being determined by reading back 

 between the site limiting curves to the 100-year point. 



Site determinations may frequently be checked by measurements 

 of dominants of the same species but of different age classes, when such 

 occur on the area. Whenever possible, determinations should be based 

 on several measurements for the same species. For many species it 

 will be unsafe to rely upon determinations based upon young trees, of 

 less than 30 or 40 years. "Giant," "dwarf," and "wolf" trees should, of 

 course, be avoided. 



The trees selected as site criteria should be forest, not open-grown 

 trees. It is possible that later studies will reveal relations between the 

 rate of height growth of isolated and forest-grown trees of the same 

 species on the same sites, by which isolated trees may be used as site 

 indicators; but until this is done the choice should be limited to dom- 

 inant trees wKich have participated in the crown cover. These may, 

 of course, be trees which have been left isolated in cuttings, and 

 when this is the case there can be little objection to their use, provided 

 their former dominance is evident from crown shape and size, size of 

 bole, intolerance of the species, etc. Even then, the determination 

 should not depend upon a single individual, but upon an average, or 

 the measurement of trees of other age classes as well. 



It is thus apparent that the adoption of this general method of site 

 classification would place a large premium upon growth data of all 

 kinds. A tremendous incentive would be given for studies of height 

 growth of all our species, with reference not only to site but also to 

 the effects of differences in the density of stocking, etc. Compared 

 with studies resulting in normal yield tables, height-growth data are 

 extremely inexpensive. For many of our species normal yield tables 

 simply cannot be had at present; even the preliminary attempts are 

 costly and require much judgment and effort. In order to make the 



