structure were at the same time brought to light; but we can scarcely
expect that the as yet unknown species will supply any link connecting
the Trochilidae with other existing families of birds. The eventual
conclusion will perhaps be that this family has come down independently
from an exceedingly remote past, and with scarcely any modification.
While within certain very narrow limits humming-birds vary more than
other families, outside of these limits they appear relatively
stationary; and, conversely, other birds exhibit least variability in
the one direction in which humming-birds vary excessively. On account of
a trivial difference in habit they have sometimes been separated in two
sub-families: the Phaethornithinae, found in shady tropical forests; and
the Trochilinae, comprising humming-birds which inhabit open sunny
places--and to this division they mostly belong. In both of these purely
arbitrary groups, however, the aerial habits and manner of feeding
poised in the air are identical, although the birds living in shady
forests, where flowers are scarce, obtain their food principally from
the under surfaces of leaves. In their procreant habits the uniformity
is also very great. In all cases the nest is small, deep, cup-shaped, or
conical, composed of soft felted materials, and lined inside with
vegetable down. The eggs are white, and never exceed two in number.
Broadly speaking, they resemble each other as closely in habits as in
structure; the greatest differences in habit in the most widely
separated genera being no greater than may be found in two wrens or
sparrows of the same genus.

This persistence of character in humming-birds, both as regards
structure and habit, seems the more remarkable when we consider their
very wide distribution over a continent so varied in its conditions, and
where they range from the lowest levels to the limit of perpetual snow
on the Andes, and from the tropics to the wintry Magellanic district;