BY THE SELECTION OF SOMATIC VARIATIONS. 55 



indicate clearly that the type in question is derived more or less 

 directly from C. hlumei. The flowers agree with the description of 

 C. hlumei. 



Coleus hlumei, as stated above, was already in cultivation in Java 

 when described by Blume. Blume (1826) suggests that the plant 

 named by him Plectranthus laciniatus may have been simply a variety, 

 as it seems to have differed largely in having laciniate leaves. Blume's 

 original description speaks of the leaves of C. hlumei as spotted above, 

 but the colored plates appearing in 1852 and 1853 show the greater 

 part of the upper surface of a soUd purphsh color. Coleus hlumei was 

 introduced into the German gardens under the name Plectranthus 

 concolor var. picta (Gartenflora, 1853, 2: 220). Only in one of Bause's 

 first lot of hybrids was C. hlumei concerned, but this species was, it is 

 stated, the seed parent of the second lot of hybrids. If this hybrid was 

 used in further hybridization work there is no record. It was the least 

 brilliantl}' colored, possessed no yellow, and sold for the sum of 5 guineas, 

 which was the lowest sum paid for any one of the 12 hybrids (Gard. 

 Chron., 3>o, 432). I have been unable to find further mention of this 

 hybrid, which was named reevesii. 



C. hlumei produced in 1868 a bud sport with the green changed to a 

 decided yellow tint. Through propagation this gave rise to the variety 

 telfordi aurea (Gard. Chron., 33, 460). 



Andre (1880) illustrates and describes 4 new varieties, which he 

 attributes to C. hlumei. Apparently all have, however, leaves with 

 cordate bases and not at all cuneate, as is the case in C. hlumei, which 

 makes his determination of doubtful validity. The same is true of the 

 type Gloire de Dijon, described and figured by Rodigas (1888). Rodi- 

 gas later (1892) notes the wide variability obtained from the seed prog- 

 eny of what was considered as C. hlumei from Chile. The 4 derived 

 types illustrated possess, however, strongly cordate leaves which make 

 the identity with C. hlumei doubtful. 



At the present time it does not appear that any pure strains of C. 

 hlumei, C. gihsonii, or C. veitchii are in cultivation. The strain used 

 in these experiments agrees most closely in regard to leaf-shape with 

 the original C. hlumei, but the variability of the seed progeny seems 

 to indicate that it is not a pure strain. The writer inquired about and 

 observed all types of Coleus available at numerous botanical gardens 

 and nurseries during a 6 weeks' trip to Germany, Holland, and Eng- 

 land during the summer of 1914. Only one plant, a plant observed 

 at the Royal Botanic Gardens at Regent's Park, London, was seen 

 which had the hlumei type of leaf. 



Coleus verschaffeliii is, however, quite generally in cultivation at the 

 present time and agrees quite closely with the type first described. 



I am especially indebted to Mr. F. J. Chittenden, of the Royal 

 Horticultural Society, for securing a statement (a letter to Mr. Chit- 

 tenden) from Mr. B. Wynne, in which he states that he spent 3 months 



