Origifi and Development of the Composites. 205 
angiospermous species found by Treub (36) in the interior of the 
island and not on the strand three years after the eruption four 
species were Composites: Wollastonia {—Wedelia) sp., Conyza 
angustifolia, C. indica and an unidentified species of Senecio. In 
the absence of other evidence it is at least possible that this 
“ Seuecio ” of Treub was Emilia sonchifolia, which is very like a 
Senecio, and which was found by Penzig (29). No Senecio was 
found by either Penzig or Ernst (15). The former found five and 
the latter eight Composite, of which Blumea balsamifera and 
Venionia cinerea were found by both and Emilia sonchifolia 
only by Penzig in 1897, fourteen years after the eruption. 
These three species are particularly interesting as they occur 
again in the list of ten species of Composite found by Gates (19) 
on Taal Island (in Bombon Lake, Luzon) about three years after 
an eruption there which denuded the island of vegetation. The re¬ 
vegetation was naturally more rapid than on Krakatau as the source 
of the seeds was about 6 km. instead of about 18 km. distant. In both 
cases a number of Composites were strand plants, evidently dispersed 
by ocean currents, but on Krakatau four species of Compositae were 
undoubtedly transported by the wind, and at least seven species on 
Taal were probably wind-borne, two ( Blumea sp. and Pterocaulon 
cylindrostachyum) occurring only on the crests of mountain ridges 
and the other five elsewhere in the interior. 
The evidence from Krakatau is, therefore, confirmed by the 
observations on Taal Island, and there remains no doubt whatever 
of the usual dispersal by wind of a number of Composite to distances 
of 4-20 miles and the occasional dispersal of pappose fruits to 
distances of over 100 miles. 
B. Experimental Wind-Dispersal. 
There is apparently no record in which wind of a known 
velocity has been used in the experimental dispersal of pappose 
fruits; all the experiments have been done on the rate of fall in 
quiet air. It is, therefore, considered advisable to give some details 
of the apparatus and method used in the present investigation. 
The apparatus and method were demonstrated to Sect. K of the 
British Association in 1916, and to the Linnean Society in the same 
year (33a). 
Apparatus. 
The source of the wind was an electric fan (Fig. 25, A) 
