— 247 - 



shinkoensis, honç^kongcnsis and speciosa, we hâve to iiieet the surprising 

 fact tliat the third mentioned species hns an epidermis widely divergent 

 from the regular Camellia-type '). T^erhaps we iiiay, as to its woody calyx, 

 refer to Tutclieria spectalnlis, formerly considered as a sister-species to 

 C- reticLilata, from which it mainly differs by its woody sepals — There is 

 indeed a good deal of probability in the assumption that such di- and 

 convergences are due to hybridization, to the separating and combining 

 of gènes 



Again, most Camellici-spedas hâve leaves with narrow bases. Only 

 two hâve definitely ovate leaves; thèse are C. Forrestii and C. Henryana, 

 both occurring in the province of Yun-nan. Moreover, a characteristic of 

 C. Forrestii, its prominent nervature, is met with in C. Henryana more or 

 less distinctly. To my thinking there was, however, sufficient reason to 

 describe the latter as a distinct species. The shape of the Forrestii fruit 

 is unknown yet; but it is a significant fact that the fruit type of the tea 

 plant, rare otherwise among the genus Camellia, is found in C. Henryana, 

 which points to a relation with tea. Now, in C. t/ieifera the trivial type 

 is a three-lobed one; nevertheless, as 1 mentioned above, in some indi- 

 viduals of the tea plant the three-lobed fruit assumes a trigonal shape 

 very suggestive of the gênerai Camellia type (as De Candolle described 

 and Wallich pictured it). 



Lastly, the coating of the ovary affords some remarkable instances of 

 the Mendelian principle of independency. Besides the case of C. Japonica, 

 having a glabrous ovary whereas the closely related C. reticulata etc. 

 possess a densely hirsute one, we find C. parvifolia and C. p«/2C/flto, being 

 the only hirsute species in the otherwise glabrous Theopsis section. On 

 the other hand, the section Thea présents one species with a glabrous 

 ovary, and one or two with a pubescent one. 



Further investigations into this genus may bring more facts of this 

 nature to light(f.i. connected with flower colour; purple in C. a/np/^x/cm/Z/s^), 

 yellow in C. hitescens and tonkinensis, notably very variable in C. japonica, 

 Sasanqiia etc.) Suffice it at présent, to hâve recognized the applicability 

 of Mendel's principles of heredity even in the genus Camellia. Of course, 

 the above considérations should be tested by experiment before far- 

 reaching conclusions were drawn from them; but such experiments are 

 likely to meet with serions difficulties, to judge from my investigations 

 into the ecology of the tea plant as described in this paper's second 

 division so that I am afraid our conceptions of the genetical interrelations 

 of the Camellia species will never exceed the province of hypothesis. In 

 this respect, however, the genus Camellia is not in a worse condition 

 than by far the greater number of systematical groups. 



') See J. KOCHS 1900, p. 608 (tab. VIII fig. 5j. 



2) Pink in C. thcifera var. rosea Mak. ? See p. 257. 



