£f-*ioo 



b + c 



where a is the lower of the two relative cover values tor a species that is 

 held in common between two sites, b is the sum of the relative covers of one 

 site, and c is the sum of the relative covers of the second site. A simi- 

 larity index is a quantitative means of determining the similarity between two 

 sites. The range of index values is from (nc species in common) to 10U (all 

 species in common with equal prominence). For this analysis, the prominence 

 value was used rather than relative cover alone, to take into account more 

 detailed data. Table 57 is a matrix of all possible comparisons. 



Based on the similarity matrix, a two-dimensional ordination was con- 

 structed of the 15 sampled belts (Beals, 1960; Fig. 131). An ordination is a 

 means of spatially separating samples based on any similar quantitative meas- 

 urement. Those belts which appear farthest apart are the least similar; those 

 that are closest are most similar. By assessing the community structure of 

 each area separately and all areas as a whole, subtle distinctions can be made 

 between different types of communities and the different structures of these 

 communities with relation to the length of time since overwash. 



From the two-dimensional ordination it is evident that s vegetatively, 

 belts X and Y (the two CF.RC areas) are grouped with belts that have recently 

 developed or overwashed, ".elts J and L (Fig. 131; group I). These belts are 

 principally dominated by Ammop-.ila breviligulata with consistent drift-line 

 subdominants. These young areas have low species richness, particularly when 

 drift-line species are discounted (Table 56). Overall biomass is high on 

 young dunes, with the highest Armophila breviligulata biomass found on accret- 

 ing d'..nes (Table 55). 



Oldest sites (belts B, C, E, and F) which have, not overwashed in the last 

 90 years are grouped broadly to the far left (Fig. 131, group II). These 

 areas have highly diversified, broad sand dunes and salt marshes. In all 

 four of these belts the dunes are 5 to 10 meters high and continuous. Belts B 

 and F are grouped close together because they both have well-developed 

 Hudsonia communities with Spartina patens and Agropyvon pungens characteristic 

 of marginally supratidal substrate. These grasslands may indicate that these 

 belts formed as a result of massive overwash which, when deflated, varied 

 gradually in elevation from low and high marsh zones to a broad transition 

 zone between marsh and dune. Supratidal grassland communities are absent from 

 young areas. 



Belt A groups with the older belts (group II) even though the shrub and 

 dune communities have only developed since 1952. The salt marsh and grassland 

 at belt A are very similar to belts B, C, E, and F. A broad Spartina patens 

 high marsh is bordered by a narrow, poorly productive Spartina. alterniflora 

 low marsh. The dune and shrub communities of belt A also do not differ 

 greatly from the other belts, indicating that in as little as 26 years dunes 

 can form and become senescent. The Arrmophila bveviligulata biomass on the 

 west side of the dune field is as depauperate as any on North Beach. The 

 shrub community at belt A is as diversified as the communities at belts B 

 or C. 



209 



V- - 



