surface of the ground, is all at once seen to be full of them, rushing
past with extraordinary velocity in a north-easterly direction. In very
oppressive weather, and when the swiftly advancing pampero brings no
moving mountains of mingled cloud and dust, and is consequently not
expected, the sudden apparition of the dragon-fly is a most welcome one,
for then an immediate burst of cold wind is confidently looked for. In
the expressive vernacular of the gauchos the large dragon-fly is called
_hijo del pampero_--son of the south-west wind.

It is clear that these great and frequent dragonfly movements are not
explicable on any current hypothesis regarding the annual migrations of
birds, the occasional migrations of butterflies, or the migrations of
some mammals, like the reindeer and buffalo of Arctic America, which,
according to Rae and other observers, perform long journeys north and
south at regular seasons, "from a sense of polarity." Neither this
hypothetical sense in animals, nor "historical memory" will account for
the dragon-fly storms, as the phenomenon of the pampas might be called,
since the insects do not pass and repass between "breeding and
subsistence areas," but all journey in a north-easterly direction; and
of the countless millions flying like thistledown before the great
pampero wind, not one solitary traveller ever returns.

The cause of the flight is probably dynamical, affecting the insects
with a sudden panic, and compelling them to rush away before the
approaching tempest. The mystery is that they should fly from the wind
before it reaches them, and yet travel in the same direction with it.
When they pass over the level, treeless country, not one insect lags
behind, or permits the wind to overtake it; but, on arriving at a wood
or large plantation they swarm into it, as if seeking shelter from some
swift-pursuing enemy, and on such occasions they sometimes remain