VEGETATIONAL ANALYSIS 



49 



In general, it may be said that small plots require less work than 

 large plots, both in the laying out and in the obtaining of data, 

 even though more small plots than large ones are needed for com- 

 plete sampling. At the same time, there is a further saving of effort 

 in that the total area sampled by small plots may usually be less 

 than that sampled by large plots and yet give comparably valuable 

 information. 



Distribution of Quadrats .—When the size, shape, and numbers 

 of quadrats have been determined, there still remains the question 

 of how they are to be placed efficiently and in such a fashion that 

 they will give representative data for the stand as a whole. If a 

 stand had a perfectly homogeneous composition, it would make no 

 difference where the sampling was done, but this is rarely, if ever, 

 true. Differences in the soil, drainage, and topography are usually 

 present and are reflected in the vegetation. These variations must 

 be fairly represented in the sample. It becomes necessary, there- 



FlG. 19. The distribution of quadrats in a stand according to three differ- 

 ent systems. (A) Random distribution as determined by Tippett's numbers. 252 * 

 (B) Spaced as widely and evenly as possible by survey and measurement. (C) 

 Distributed evenly along lines run by compass or sighting; spacino- deter- 

 mined by pacing. 



fore, to distribute the quadrats throughout the stand, and a plan 

 that will eliminate the human factor in placing the individual plots 

 is desirable. 



The statistician prefers a sampling system that gives him data 

 obtained at random. 216 This demands a division of the entire stand 

 into possible sampling areas and then a selection of actual sampling 

 areas determined strictly by chance. Under such conditions, the 



