1920] SARGENT, NOTES ON NORTH AMERICAN TREES. VI 245 



lected by Mr. Wilson from several large trees, among which I find only a 

 single leaf with one lateral lobe, the leaves of older trees are almost invari- 

 ably entire, but they are partly three-lobed on his specimens of young 

 plants. The only fruit collected is smaller than that of the Chinese species 



and measures 6 mm. in diameter. 



The Forniosan species occurs only in scattered individuals on the moun- 

 tains Arisan and Kandaisan and possibly elsewhere on the island, while 

 each of the other two species has a much wider distribution and is much 

 more common within its range. All the species, however, are only relics 

 as compared with the range of the genus during the Tertiary period, when 

 it was distributed apparently all over the northern hemisphere. 



NOTES ON NORTH AMERICAN TREES. VI 



1 



C. S. Sargent 



Hamamelis 



The different species of Hamamelis cannot be distinguished by good mor- 

 phological characters, the structure of the flowers, fruits and seeds being the 

 same in them all, and in View of the superficial botanical surveys on which 

 the earlier publications on the plants of the southern states were based it is 

 not surprising that all the Witch Hazels in North America have usually 

 been referred to a single species, Ilam^amelis rirginiana L. Walter, how- 

 ever, in 1788 mentioned without descriptions three species, dioica, monoica 

 and androgijna, and in 1814 Pursh described his 77. macropJujUa based on a 

 specimen collected by Lyon on river banks in the w^estern part of Georgia 

 and chiefly distinguished by the tubercles on the lower surface of the leaves. 

 After Nuttall in 1818 and Elliott n 1821 no American author has considered 

 Pursh's plant which has been referred without comment to H, virginiana. 

 In 1911 I described as 77. vernalis a small winter-flowering shrub spreading 

 into thickets by stolons and con.mon on the gravelly banks of streams in 

 southern Missouri and now known to extend through Arkansas Into eastern 

 Oklahoma and distinguished from 77. virginiana by the red color of the 

 inner surface of the calyx-lobes and by the sha])e of the leaves which are 

 often round at apex, more symmetrical at base, less coarsely lo})ulate and 

 pale and often glaucous below. At one time I believed that the Witch 

 Hazel of Louisiana could be referred to this species, but the leaves of all the 

 specimens of Hamamelis which I have seen from the Gulf States are covered 

 with tubercles. These are the enlarged bases of the short stellate hairs 

 which cover the young leaves and are more prominent on the lower than on 

 the upper surface. Such tubercles are not common In Hamamelis, and have 

 not been found on the leaves of any other species except occasionally on 



one of the Japanese species. These tubercles, the smaller flowers which 

 open from December to February with a calyx only 5 mm. across the flat- 

 tened lobes and pale yellow sometimes streaked with red, petals only 8 mm. 



1 For part. V, see p. 61. 



