THE STRUCTURAL CHARACTERISTICS OF THE COMMUNITY 39 



consideration, according to Kylin (1926). Hyperdispersion works the 

 reverse; the greater it is the more heterogeneous appears the vegetation. 



The homogeneity problem, brought forward by Nordhagen (1922), 

 has been theoretically treated from the mathematical side (KyHn, 

 1926; Romell, 1926), without being fully cleared up. 



In phytosociological practice, dominance and homogeneity are 

 closely related. Areas in which one or a few species predominate, with 

 even distribution, are considered homogeneous, since the more or less 

 unevenly distributed but scarcer species are mostly hidden by the 

 dominants. Regions with few species, with dense plant covering made 

 up of crowded perennials (subarctic, arctic), present much more homo- 

 geneous communities than regions rich in species but with open vegeta- 

 tion and many therophytes (middle and south European mountains, 

 subtropics) (Fig. 17). 



This may explain the different concept and definition of the associa- 

 tion held by many northern investigators. They regard fidelity as of 

 less importance than homogeneity in the characterization of an associa- 

 tion, the latter being more apparent in the north. 



Frequency. — Frequency graphs, as determined by the Raunkiaer 

 method, may be considered as an approximate expression of the 



/ 2 3 



Fig. 18. — Sample plots for the determination of frequency, showing various arrange- 

 ments, in a stand (A). 



homogeneity of a stand. Raunkiaer (1913, 1918) based his studies of 

 "frequency" upon small quadrats or circles of 0.1, 0.5, or 1 sq. m., 

 taking 25, 50, or 100 at random in the stand investigated. The Hst of 

 species of each small area was noted, without counting individuals. 

 Determinations of frequency (under the confusing name of con- 

 stancy) recently have been made with fewer (10 to 20) and larger 

 (1 to 4 sq. m.) quadrats (Fig. 18, 1, 2). The test areas as in the deter- 

 mination of density are arranged in a definite order (Clements, 1905; 

 Lagerberg, 1915; Allorge, 1925) or are distributed at random in a 

 uniform community (Raunkiaer 1918) (Fig. 18, 3). The result may be 

 expressed in a diagram, distributed in 5- or 10-frequency classes. 



