5 
Botanists more nen ~ narra classification to suggest such modi- 
fications as may improve the groups which are now defective 
In the first he Reber, I have followed Bartling, a some 
ae and the addition of Calceolaria and Jovellana, which 
appe e to connect a with Alonsoa. It remains 
compose. of the following gene 
m is a very natural ila genus if taken in conjunction 
with Celi — which it is artificially separated by the presence of 
the fifth s n. The German species have been well distributed 
by Reichenbach (Fl. Germ. Excurs. 383) into three sections, Thap- 
sus, Lychnitis, and Bilattaria, but the separation of the species is 
difficult. Romer and Schultes enumerate 69 ; Pp aecear 8e has 35 
German ones; Tenore 24, in his Neapolitan floras ; Sprengel re- 
the most part incorrect, yet the great extent to wade the same spe- 
cies will vary in > texture and quantity of simi and the form of 
the pu cajun as. acciden: ie, hy- 
bride, or Nelidinee an a the number of legitimate species even 
below that admitted by Sprenge 
Celsia includes several plants, differing more in habit from each 
other than from Verbascum and Alonsoa, but artificially distinct 
from both. It may be naturally divided into three sections, as 
follows :— 
Sect. 1. Pseudothapsus. Herb. Stamina 2 longiora glabra, an- 
theris oblongis adnatis, 2 breviora barbata, antheris medifixis bilocu- 
us loculis confluentibus. C. cretica, betonicefolia, &¢. 
Siests 2. Arcturus. Herbe. Stamina 2 longiora glabra, breviora 
leviter barbata, antheris omnibus medifixis bilocularibus loculis con- 
omandeliana. 
Sect. 3. Nefflea. Suffrutices. Filamenta omnia barbata. Antheree 
omnes terminales ecietieres valvuli s 2 ovatis acutis dehiscentes. 
Alonsoa is so very different from He ris in the str ructure of the 
olaria, a very extensive genus, r cane as much as any 
the laboute of a monographist, for the distribution of the species 
