144 



R. H. RICHENS 



An examination of Fig. 3 indicates that the elms collected fall into a number of 

 groups and subgroups. These are listed in Table 2, together with the minimum and 

 maximum values for the five characteristics studied in each assemblage; a note is added 

 on the probable relation of each assemblage to the elm species and varieties recognised 

 by Tutin (1952). A hyphen is used to indicate intermediate characteristics, not necessarily 

 hybridity. 



AXONOMIC ANALYSIS 



It will be assumed that the groups and subgroups listed in Table 2 are genetically 

 diverse entities. Without transplantation experiments it is impossible to rule out the 

 possibility that some of the differences observed are environmental rather than genetic, 

 but it seems improbable that environmental effects are a major source of variation because 

 (a) within a single tree or clonal stand, variation in the characteristics investigated was 

 usually very small, (b) no correlation was detected between any particular type of variation 

 and habitat, and (c) striking differences were frequently noted between adjacent clones 

 in a single hedgerow. 



Coming now to the groups recognised, it is evident that group I is a highly variable 

 assemblage showing smooth intergradations between the extremes. The subgroups are 

 interconnected thus : 



c 



h 



\a is a large assemblage characterised by a low tooth number and low obversity; it 

 includes both smooth and scabrous leaved types. It is closely connected to Ig, a scabrous- 

 leaved type with notably broad leaves, to the smooth-leaved subgroup lb with rather 

 long petioles but low asymmetry, and to Ih, characterised by its high tooth number. 



A very homogeneous assemblage is constituted by Id, with both smooth and scabrous 

 leaved representatives; the tooth number is high, the obversity low. la and Id form a 

 ring with le and If. le resembles Id in its low obversity, while If has rather narrow leaves 

 with shortish petioles. 



Ic stands rather apart from the others. It has broad leaves with very long petioles 

 and highly asymmetrical bases. 



It is clear that I as a whole, excluding perhaps the scabrous members, comes within 

 U. carpinifolia Gled. as interpreted, for instance, by Rehder (1940). In the rest of this 

 paper, U. carpinifolia will denote this broad conception of the species; U. carpinifolia 

 Gled. emend. Melville (1946) will be referred to as U. carpinifolia sens. str. The scabrous- 

 leaved types are intermediate between U. carpinifolia and U. glabra Huds. Attempts, 

 however, to equate the subgroups with Melville's microspecies are difficult. The type 

 description of U. diversifolia Melville (1939) applies to some members of la, though the 

 coexistence of asymmetrical and symmetrical leaves on dwarf shoots from the same tree 

 was never observed. lb and Ih also come fairly near to U. diversifolia. It should be 

 pointed out, however, that the straight- sided leaf margin making a right-angle turn into 

 the petiole, regarded by Melville (1946) as diagnostic of U. carpinifolia sens, str., occurs 

 in some members of la and in Ih. Ic has a remote affinity to U. coritana var. media Melville 



