10 CIRCULAR 3 3 3, U.S. DEPARTMENT OF AGRICULTURE 



10 links, or one one-thousandth of an acre). If seedlings and other 

 small vegetation are to be mapped in detail, it may be much more 

 convenient to use a number of half-milacre subplots (10 by 5 links). 

 The smaller size is particularly advantageous where frequent ex- 

 aminations are necessary, and is highly desirable on steep slopes. 



In studying herbaceous vegetation, as in ecological investigations, 

 quadrats 1 or 2 feet square may well be used. 



Sample plots should be compact. The shorter the perimeter, the 

 less the problems of border influence and line trees ; and the chance 

 of homogeneity is greater for compact plots. On the other hand, 

 the danger of inadequate sampling is greater in the case of square 

 plots than in that of strip or line plots ; and under some conditions, 

 as in very open stands on cut-over areas or in semiarid regions, long 

 narrow strip plots are usually preferable. 



In most cases the length of a plot should not be more than 2.5 

 or 3 times its width. 



Whenever uniformity of conditions permits a choice as to shape, 

 the preference should go to rectangular plots the sides of which are 

 even units of measure — especially if the stand is to be treated. Such 

 plots are readily surveyed, mapped, and divided, and their bound- 

 aries are easily marked and relocated. 



Under some conditions irregularity in shape is unavoidable, as 

 where the forest type to be sampled occupies only a very narrow belt 

 such as the bottom of a canyon. In such cases the plots should 

 have as few sides as possible; and reentrant angles, and any angles 

 less than 60°, should be avoided. Where the boundaries of the type 

 area are so irregular as to be extremely difficult to survey on the 

 ground, it is advisable to lay out a rectangular plot overlapping the 

 area and plot on the map the exact portion of the plot on which the 

 type occurs, either by traverse or by strip surveys. This portion 

 can then be planimetered and the tree data taken and recorded 

 accordingly. 



Circular plots are easy to lay out, and since the perimeter of a 

 circle is smaller in proportion to area than that of any other figure 

 they are subject to the least possible border influence. To mark a 

 circular boundary accurately, however, requires much time, espe- 

 cially in young stands. On circular plots there is greater danger 

 of inaccuracy in all the operations of plot establishment than on rec- 

 tangular plots, and satisfactory division is impossible; in general, 

 therefore, circular shape is not recommended for permanent sample 

 plots. 



It cannot be emphasized too strongly that a sample plot's size 

 and shape, as such, are of much less consequence than its representa- 

 tive character. 



The desirable size and shape of subplots, and the desirable number 

 per acre or plot, can best be determined through a special statistical 

 study of the forest type and of plot conditions. For this purpose, 

 one or several areas representing stand conditions comparable to 

 those that can be expected to exist on the plots following treatment 

 should be divided into a number of small subplots that can be com- 

 bined into units of any desired size. The standard errors derived 

 from these combinations will indicate what size is most desirable 

 for the quadrats {32, 33). 



