through the interstices of the cinders, 
where they are protected from agen- 
cies, primarily the wind, that might 
otherwise carry them farther. Also, un- 
der most conditions, seeds that remain 
on the surface lie within a laminar 
boundary layer of still air, and only a 
very high wind can move such small ob- 
jects (0.6 mm long and 0.4 mm wide). 
Consequently, I reasoned that my traps 
would not artificially restrict the dis- 
persal of seeds. 
I found that trains are much more ef- 
fective than the wind in dispersing 
seeds of dwarf snapdragon for long dis- 
tances. Despite the periods of strong to 
gale-force winds that occurred during 
some of the wind-dispersal experi- 
ments, in no case did I find seeds far- 
ther than eight inches from the parent 
plant; most landed much closer. By 
contrast, a locomotive and eight box- 
cars, traveling at about twenty-five 
miles per hour, dispersed half the seeds 
within eight inches of the parent plant 
and the rest up to five feet away. Some 
seeds might be carried even farther if 
they stick to the wheels or underside of 
a train and are dislodged later; the fast- 
er the speed of the train, the farther 
along the tracks it will carry seeds. 
Mark Widrlechner of the University of 
Minnesota has suggested that parts of 
mature plants may break off and be car- 
ried along the tracks like tumbleweed, 
dispersing seeds as they go. Short-range 
dispersal of seeds by the wind and pass- 
ing trains may allow dwarf snapdragon 
to exploit locally favorable areas, and 
trains may also facilitate the long-range 
spread of the species along the tracks to 
new areas. 
These findings fit well with the ob- 
served distribution of the dwarf snap- 
dragon in the United States. In the 
Mediterranean region, the species 
spread to newly built railroads from 
nearby fields and waste areas. In the 
l Si 
62 
