STUDIES ON ULMUS : I 



145 



(1949), but the leaves are narrower, and the petioles longer, than in the type description 

 of this variety. Id, le, and If do not correspond closely with the type descriptions of any 

 of Melville's segregates. 



The second largest group, II, is quite sharply differentiated from I. The principal 

 subgroup, Ila, has ovate to orbicular leaves with a low tooth number. The leaves are 

 usually scabrous but exceptions occur. Two samples with a higher tooth number are 

 relegated to lib; these forms have smooth leaves. 



The scabrous members of Ila correspond exactly with U. procera Salisb. The 

 smooth-leaved members of this subgroup are more of a problem. Their relationship 

 with U. procera is obviously close. On the other hand, they might be regarded as extreme 

 forms of U. carpinifolia; they bear some resemblance to the microspecies U. coritana but 

 do not correspond at all closely with either var. rotundifolia Melville, which has fewer 

 teeth and more symmetrical leaves, or with var. media, which has longer petioles and 

 more highly asymmetrical leaves. Another possibility is to regard these forms as inter- 

 mediate between U. procera and U. carpinifolia.- lib can probably be classed as a member 

 of U. carpinifolia, since the high tooth number as well as the smooth leaf surface differentiate 

 it fairly sharply from U. procera. It corresponds quite closely with U. coritana var. media. 



Ill includes a set of narrow-leaved elms with high tooth numbers. They differ 

 from 1/ in having longer petioles. Forms with smooth and with scabrous leaves both 

 occur. The group, in spite of its scabrous tendency, can probably be classified under 

 U. carpinifolia. The allocation to a microspecies is less certain. The nearest is U. 

 carpinifolia sens, str., but only some members of III show the straight leaf and right 

 angle turn into the petiole characteristic of this segregate. 



rV differs from la in the longer petiole, greater asymmetry and higher obversity. It 

 clearly comes under U. carpinifolia, but does not correspond well with any of the micro- 

 species. Sample 26 has the characteristic leaf shape of U. carpinifolia sens. str. but in 

 other respects recedes considerably from the type description of this segregate. 



Five samples, 52, 70, 90, 91 and 92, though all significantly different from each 

 other and from II, the nearest group, resemble each other in being smooth and broad- 

 leaved elms with relatively few teeth, markedly long petioles and low asymmetry. They 

 are associated into group V. The two samples of Va differ from Vt> in their rather shorter 

 petioles and more pronounced basal asymmetry. Sample 91, the sole representative 

 of Vc, has broader leaves than either of the other subgroups and very low obversity. The 

 group as a whole comes within U. carpinifolia, but is far more sharply separated from 

 the other members of this species than any of the segregates hitherto considered. As 

 will be seen below, this group has a distinctive geographical distribution. Va is fairly 

 close to U. coritana var. rotundifolia though having a rather longer petiole. The other 

 two subgroups do not correspond closely with any of the previously described segregates. 



The two scabrous-leaved samples 07 and 85h, though differing significantly in tooth 

 number, agree in having broad leaves with very shortpetioles, low asymmetry and low 

 obversity. They are typical representatives of U. glabra, and are consigned here to 

 group VI. 



Three samples were collected which were intermediate between VI and the smooth- 

 leaved elms of I. Sample 04 has broad scabrous leaves with an intermediate tooth number, 

 the petiole being significantly longer than in VI. It approaches close to sample 64 of 

 Id, but the leaves are slightly more symmetrical. Samples 15 and 61fc» resemble VI in 

 their high tooth number, but in other respects, including the smooth leaf surface, come 

 fairly close to Id. These last two samples differ considerably from each other, 15 having 

 longer petioles and much lower obversity than 61b; they are therefore relegated to separate 

 subgroups, Vila and Vllb, respectively, of a single group. This group and 04 are clearly 



