SAMPLE PLOTS 1ST SILVICULTURAL RESEARCH 9 



An experiment cannot be considered complete, therefore, until plots 

 have been established, and measurements made, during a sufficient 

 number of different years to show the significance of annual differ- 

 ences in weather. 



Under some circumstances replications, while increasing the area 

 required in a study, may fail to give any greater accuracy. In 

 many field-crop investigations, this difficulty has been largely over- 

 come by confining the comparisons from which both errors and esti- 

 mates are derived to limited areas, through various plot arrange- 

 ments. Such schemes, of which the Latin square (28) is typical, 

 should be used more extensively in forest research. No one system 

 of plot arrangement has yet been proved the best. In fact, in forest 

 research so few studies of arrangement have been made that the 

 field can be considered practically unexplored. A wide variety of 

 possibilities are open. Only by practical tests largely paralleling 

 those made in field-crop investigations can the best, most economical, 

 and most practical method be determined. 



For subplots laid out for the purpose of studying quantitatively 

 or qualitatively the reproduction on a sample plot, the proper num- 

 ber and distribution can be determined statistically (68). Here the 

 plot is the area being sampled. 



SIZE AND SHAPE OF PLOTS AND SUBPLOTS 



How small a sample plot will include enough trees to furnish aver- 

 ages of the required representative character depends on the density 

 of the stand, variations as to diameter and height, the number of 

 different crown classes, the number of different species, etc. The 

 number of trees should be sufficient to give a reasonably reliable dis- 

 tribution curve for such factors as diameter and height. Since den- 

 sity changes with age, a plot area that is ample for a young stand 

 may be much too small when the stand becomes older. 



As a general rule, a plot should be of such size that at the end 

 of the experiment it will contain at least 100 trees of the predominat- 

 ing age class. Under ordinary circumstances it should rarely be 

 large enough to contain more than 1,000 such trees at the end of the 

 experiment, if only because of the work involved. Where the small 

 size of suitable homogeneous stands subject to uniform conditions 

 makes it impossible to establish plots of the size recommended, 

 smaller plots are permissible. 



Where conditions permit, the area of each plot should be an even 

 multiple or simple fraction of an acre. Simplicity of area conver- 

 sion is an advantage throughout the computation and interpretation 

 of plot data. 



A rectangular subplot may be of any desired length. If data are 

 recorded separately for even units of length, new subplot divisions 

 can be decided upon in the office after the data are examined-. Rec- 

 ommended units of length and width are the chain and the one-tenth 

 chain. 



Subplots in small reproduction that is to be studied intensively 

 should be small enough so that a man standing outside a subplot 

 can reach any part of it. A convenient unit is the milacre (10 by 



