— 276 — 



stigmas Vz o^ ^^^ style), sometimes split down to the ovary, length about 

 1 cM. Fruits glabrous, 2-3 cM. in diameter, 1-4-lobed according as seeds 

 hâve set in one or more cells, containing 1-6 seeds, 1-3 seeds in each 

 cell. Seed dark brown, 1-2 cM. in diameter, spherical or mutually flattened, 

 almost smooth. — 



From the détermination table given in the 4*'i chapter it is sufficiently 

 clear in what respects the tea plant as defined above differs from the 

 other Camellia-spedes. We might epitomize thèse characteristics of the 

 tea plant in the foUowing points: 1. the stalked, drooping flowers; 2. the 

 permanent calyx, glabrous as a raie; 3. the white petals, not emarginate,as 

 a raie glabrous; 4. the stamens which are cohering only at their base and 

 glabrous; 5. the hirsute ovary and the hairless style; 6. the 3-4-lobed fruit. 



Thèse are the main characteristics in common to ail forms and shades 

 of the tea plant. We might now deal with the variations to be found 

 within the limits of this species, and with the subspecies or varieties 

 which they suggest to us. But, having had too little opportunity to study 

 the Chinese and Shan tea plants, I do not feel disposed to give a rigid 

 System of tea varieties now. This much 1 can pronounce, that 1 would 

 not retain any of the four varieties of Watt, though his Indian races may 

 be left untouched. On the other hand, at the close of the 2"'' chapter I 

 cursorily made an attempt to group the tea forms in four new divisions, 

 thus: 1. Manipur, Cachar, etc. (large-leaved India indigenous), 2. (Assam?) 

 Burma and Siam, with Tong-king? (small-leaved India), 3. macrophylla 

 from Yun-nan and Sze-chuen (large-leaved China), 4. bohea + viridisirom 

 Eastern China (small-leaved forms). So, roughly speaking, we might perhaps 

 distinguish an indica, burmensis, macrophylla and sinensis group; but I 

 cannot yet undertake to define any of thèse groups in détail. ') 



A similar classification, treating ail varieties, as it appears to me, on 

 equal terms, is préférable to a System where again and again the Chinese 

 forms are endowed with a prédominant position, and such merely because 

 LiNNAEUS knew them first by chance. „Auf Grund der Diagnosen van Linné", 

 as KOCHS 2) rightly observes, „wird es jetzt wohl niemandem mehr môglich 

 ,,sein, aus ail den Varietâten Bohea oder viridis mit Bestimmtheit heraus- 

 „zufinden." But why, then, do ail authors attempt to retain them and attribute 

 new characters to them — a hopeless task that inevitably must resuit in 

 arbitrariness? They were tempted to describe the „levissimae varietates", 

 an employment from which Linné himself had warned them off; they ought 

 to hâve formed larger groups while working for systematical purposes. 



As it is, KoCHS 3) retains the names 5o/zea and v/Wc^/s for two varieties 

 which are according to his diagnoses „vollkommen" distinct, ôz/Zw/z/c/z /zav^ 



') In the subséquent division of my paper, dealing with the application of statistical 

 methods of description on sélection, I hâve tried to form some morphological groups 

 on account of characters derived from the living plant. 



*) J. KoCHS 1900, p. 600. 



') J. KoCHS 1900, p. 601. 



