STUDIES ON ULMUS: I 



149 



but since the scarp is narrow, this feature has hardly affected the sampHng. The Breckland 

 (for example minor square 77) provided no elms, and elm is sparse on the limestone 

 outcrop in the extreme north-east part of the area. 



The most striking result of the survey in respect of the distribution of the various 

 groups is the discovery that U. procera, in the older hedges, does not extend eastward 

 of a line running approximately north-south from Peterborough to Walth^m (cf. Fig. 1). 

 This is especially noteworthy in view of the fact that U. procera is frequently planted in 

 later hedges and grows well east of this line. It is of interest to note that this line, to 

 anticipate the results of further work, bends round sharply south of our area, and runs 

 eastward, parallel with the Thames, reaching the sea near Shoeburyness. It reappears 

 in east Kent, near Herne Bay, and runs thence in a south-westerly direction, leaving a 

 small area in south-east Kent where U. carpinifolia is once again the predominant elm. 



U. glabra occurs only sporadically throughout the area, suggesting a relict distribu- 

 tion. It is fairly frequent on the oolitic limestone outcrop, for example, in minor square 

 07, but this formation only just comes into the area. It is possible, however, that the 

 widespread scabrous-leaved types, here considered under U, carpinifolia, have resulted 

 from introgression from this species. 



U. plotii, or rather the related sample 06, was only encountered once. This is not 

 surprising since this species is known to be characteristic of the country to the west of 

 Stamford. It appears to reach its eastern limits around Kimbolton. 



U. carpinifolia occurs throughout the whole area. It penetrates beyond the eastern 

 boundary of U. procera, though with reduced frequency, but how far it extends westward 

 has not yet been ascertained. It has been collected along parish boundary hedges as 

 far west as Bletchley. 



What is of more interest is the distribution of some of its constituent groups and 

 subgroups. Thus Id, with leaves of medium breadth and high tooth number, is charac- 

 teristic of the country between Royston and Clare, while the elms with very long petioles 

 and highly asymmetrical leaves that constitute Ic have only been collected around Thetford. 

 As might be expected, some of the smooth-leaved representatives of Ila, the subgroup 

 which includes typical U. procera, are found where U. carpinifolia and U. procera overlap, 

 for example, 24a. Samples 56 and 57, however, which also belong to this subgroup, 

 were collected in hedges adjoining the fens. 



Ill, the group of narrow-leaved elms with a high tooth number, is fairly widespread, 

 but especially frequent in Huntingdonshire and the adjacent parts of Cambridgeshire, 

 e.g. 16, 18, 25, 35 and 48. IV was only collected in Cambridgeshire. 



With the broad-leaved elms in V, a totally different area is concerned. Va was 

 only observed in the coastal region of Essex. Sample 90 of Vb grows alongside the Black- 

 water estuary, while 52, which is the least typical representative of the group, was the 

 only inland specimen. Sample 91 of Vc also came from near the Blackwater estuary. V 

 extends south beyond our area to the estuary of the Crouch, but is then replaced by 

 U. procera. South of the Thames, it has not been found. U. procera occurs along the 

 north Kentish coast as far east as Herne Bay, while in Thanet some very small-leaved 

 elms, most nearly resembling some of the elms of the Lincolnshire limestone, appear. It is 

 not known how far north of Colchester V extends, but the elms around Walton-on-the 

 Naze and along the Norfolk coast are totally different. This group, then, which is very 

 distinct from all other elms, seems to be quite narrowly localised along the Essex coast. 



The other groups of U. carpinifolia, VIII, IX and X, are only known from single 

 collections. It is possible that these, especially the very distinctive sample 32, are single 

 extremely localised clones, but more intensive sampling would be necessary to confirm 

 this. 



