8 4 
James Small. 
tendency to reduction towards type N which appears rarely in the 
Arctotidinae is more developed in this sub-tribe, appearing in three 
genera. 
The Gundeliinae is a small group with the pappus of type F, 
connate at the base, or the reduced form, type N. 
Anthemidece. The pappus here is usually reduced to the 
fimbriato-coroniform type when present, but in a large proportion 
of the genera it is entirely absent. In a few genera, such as 
Athanasia, Lepidostephium, Allardiu, Isoetopsis, and Mnrasmodes, 
the evolution of type N has proceeded only as far as type F, the 
paleaceous setae being connate at the base and quite short in 
several of the above genera. The next step is seen in other genera, 
such as Anacyclns, Anthemis, Peutzia and Crossostephium, where the 
pappus in some species is very short and the fimbriato-coroniform 
type is more common. In many of these genera and others of the 
tribe type Qb occurs, e.g., some species of Chrysanthemum, Cotula, 
Pentzia and Anthemis ; or the pappus may be type N, or type Qb, or 
absent in different species of the same genus. 
Interesting transition forms occur in the various species of the 
anomalous genus, Oedera, types N and F occur in addition to a 
special form somewhat like type K but with the paleae fused by 
their margins to give a deep, membraneous tube, fimbriate at the 
top. 
In only one genus, Cancrina, are the paleae few in number, 
type M, and in some species of the most closely allied genus, 
Allardia, setae more or less flattened, but closely approximating the 
typical pappus of the Senecioneae, occur. Other species of Allardia 
show these setae developed into type F, with the length reduced 
and the setae connate at the base ; the apex of the setae may be 
simple or plumose. In Allardia and Cancrina, therefore, we get 
a complete series of transition forms leading from type A to types 
N and M. 
As Oedera and Allardia are in separate sub-tribes we again get 
no indication of the relative position of the Anthemidinae and 
Chrysanthemidinae and it becomes probable that the Anthemidese, 
like the Cichorieae, is sub-divided on an artificial basis. 
Inulece. The Gnaphaliinae is taken as the primitive sub-tribe 
and the Helichryseae as the primitive group within that sub-tribe, 
the Eu-gnaphalieae being derived. This arrangement is confirmed 
by the structure of the pappus. Helichrysutn and Gnaphalium are 
the chief genera ; the former has one row, or rarely several rows, 
