bulging eyes. Another attaches similar tokens to the
tongues of butterflies, while the cypripedium speeds its parting guest
with a sticking-plaster smeared all over its back. And so we might
continue almost indefinitely. From the stand-point of frivolous human
etiquette we smile, perhaps, at customs apparently so whimsical and
unusual, forgetting that such a smile may partake somewhat of
irreverence. For what are they all but the divinely imposed conditions
of interassociation? say, rather, interdependence, between the flower
and the insect, which is its ordained companion, its faithful messenger,
often its sole sponsor--the meadows murmuring with an intricate and
eloquent system of intercommunings beside which the most inextricable
tangle of metropolitan electrical currents is not a circumstance. What a
storied fabric were this murmurous tangle woven day by day, could each
one of these insect messengers, like the spider, leave its visible trail
behind it!

As a rule, these blossom ceremonies are of the briefest description.
Occasionally, however, as in the cypripedium and in certain of the
arums, or "jack-in-the-pulpit," and aristolochias, the welcome becomes
somewhat aggressive, the guest being forcibly detained awhile after tea,
or, as in the case of our milkweed, occasionally entrapped for life.

[Illustration]

From this companionable point of view let us now look again at the
strange curved stamen of the sage. Why this peculiar formation of the
long curved arm pivoted on its stalk? Considered in the abstract, it can
have no possible meaning; but taken in association with the insect to
which it is shaped, how perfect is its adaptation, how instantly
intelligible it becomes! Every one is familiar with the sage of the
country garden, its lavender flowers arranged in whorls in a long
cluster at the tips of the stems. One of these flowers, a young one from
the top of the cluster, is shown at A (Fig. 4), in section, the long
thread-like pistil starting from the ovary, and curving upward beneath
the arch of the flower, with its forked stigma barely protruding (B).
There are two of the queer stamens, one on each side of the opening of
the blossom, and situated as shown, their anthers concealed in the hood
above, and only their lower extremity appears below, the minute growth
near it being one of the rudiments of two former stamens which have
become aborted. If we take a flower from the lower portion of the
cluster (D)