Quadrat frequency and species' standing crop data were collected 

 for all communities except the sound-side shrub community. For the 

 latter type, frequency for all species and rooted stem density were 

 determined. Standing crop was determined in grams of aboveground 

 oven-dried tissue. Each quadrat was clipped, separated by species, 

 and individual species oven-dried to constant weight at 105° Celsius. 

 Five 1- by 1-meter plots and 330, 0.2- by 0. 2-meter plots were examined 

 and clipped. 



c. Vegetational Mapping . Three overflights of the study area were 

 made at different seasons. Infrared aerial photos obtained were used 

 with the ground truth from the nine previously described transects 



and the quadrat data to produce a scale vegetative map of the study 

 area. The accuracy of this map is better than 90 percent. 



d. Ordination of Stands . The 11 community types were arranged 

 in an ordination model according to the method of Bray and Curtis 

 (1957). In this method each community's frequency values were summed. 

 Each individual species frequency in the community was divided by the 

 total for all species and the result multiplied by 100 to yield 

 relative frequency expressed as a percentage. The relative frequency 

 values were then used to compare the species composition of each 

 community with the other communities, using the Index of Similarity 

 (IS), IS = 2w/(a + b) (Bray and Curtis, 1957). The IS values were 

 then subtracted from 100 to yield an Index of Dissimilarity (ID), 



ID = 100 - IS, and used to locate the communities along an axis by 

 means of Beals' (1960) adaptation of the Pythagorean Theorem, 



X = - — ^ ^^l " ^^^^ > and the technique of Bray and Curtis (1957). 



The usual procedure for axis extraction is to sum the ID values for 

 each vegetational unit. The stand with the highest sum is deemed the 

 end of an axis; the opposite end of this axis is the stand least like 

 it (i,eo, having the highest ID in relation to it). The units are 

 represented as points separated by a scale distance equal to the ID 

 value. All other points are then located between the end points. 



In this study, however, one stand had ID values of 100 with five 

 other stands, and a second stand had the same relationship with four 

 other stands. To produce a model of acceptable geometry it was 

 necessary to modify the standard axis extraction technique in the 

 following manner. The first end point of the X axis was the community 

 having the highest ID sum. The second end point was the stand with 

 the highest ID value in relation to it, excluding those stands with 

 values of 100. The Y axis was selected in a similar fashion from a 

 matrix of ID values composed of the remaining nine communities. The 

 Z axis was selected using a matrix from the remaining seven stands. 



e. Permanent Quadrats . Three permanent 5- by 5-meter quadrats 

 were located in each of the nine defined community types and three 



19 



