12 JOURNAI^ OF I'ORFSTRY 



These curves were found to be supported by. 85 per cent of the plots 

 when the three class series was used and by 68 per cent for the six- 

 class series." In most cases the plots excluded showed low cubic 

 yields, and were regarded as under-stocked. There were very few 

 cases in which the yield indicated a higher site class than that indi- 

 cated by the rate of height growth. While these exclusions were 

 to some extent arbitrary, they were thought justified in view of the 

 relatively heavy per cent of plots supporting the curves. The plots 

 represent eight composition classes, based upon the proportion by 

 cubic volume of the predominant species ; chestnut under 50 per cent 

 (50 plots), chestnut 50 per cent and over (87), chestnut oak (62), 

 white oak (51), black oak (45), scarlet oak (53), red oak (11), yellow 

 poplar (8), and miscellaneous (3). There was no important difference 

 as between these classes in regard to the proportion of plats excluded. 

 For the white and scarlet oak classes, 39 plots, or 37.5 per cent of the 

 total fell too low in cubic yield to coincide in site class as determined 

 for both volume and height, according to the six class series." 



Groivth of Irregular Stands. — General yield tables for uneven-ageci 

 and understocked stands are, of course, out of the question. A number 

 of methods for determining the yields of irregular stands have been 

 suggested,* but so far they have not proved successful. The working 



<5 It is interesting to note that the proportion supporting the table is consid- 

 erably greater than in the case of the yield tables for second-growth hardwoods 

 in Connecticut (Forest Service Bulletin 96) in which all plots deviating by 

 more than 10 per cent above or below the average basal area curve for a 

 given site were excluded. 



'' The possibility of using these tables for southern upland hardwoods outside 

 of the Southern Appalachian region was suggested, and the chestnut and oak 

 plots used for the Connecticut yield tables (Bulletin 96) were accordingly 

 classified by height on the site-class curves, and their cubic volumes plotted 

 on the yield curves just described. Only 55 per cent of the plots showed 

 a site coincidence by both height and cubic yield, most of those which failed 

 to coincide falling low in volume. It is quite likely that the use of chestnut 

 alone as a site determinant in the Connecticut study is responsible for the 

 failure of many of the plots to measure up in yield. If an average between the 

 dominant heights of several species had been taken, more of the plots might 

 have been found to fall into lower site cla,sses, consistent with their cubic 

 yields. Difference in rate of growth between Connecticut and Maryland maj 

 of course be attributed to climate, but it is also possible that a further com- 

 parison of the northern and southern plots, especially on the basis of total 

 basal area or of number of trees per acre and their average diameter, would 

 show that the discrepancy is due not so much to climate as to differences 

 in the field technique, in such directions as the basis used in deciding whether 

 plots were or were not fully stocked. It remains that a large proportion of 

 the Connecticut plots do fall in with the Southern Appalachian site classification 

 as to both height and volume growth, and comparisons are thereby facilitated, 

 of methods for determining the yields of irregular stands have been 

 ® See, for example, Proc. Soc. Amer. Foresters, Vol. 9, No. 2, April, 1914 ;. 

 articles by Chapman, Ashe, Woolsey, and Moore. 



