to which she was subjected. She was often dragged away with
her back on the ground and her legs coiled up, apparently helpless. If
all the soldiers had been treated in this way, it would not have been so
remarkable, but so far as I could see the rest were allowed to remain,
going in and out of the nest as if taking a survey of their new
surroundings.

For five months I had these ants under almost constant observation, and
yet I was unable to make out the true position of the soldiers in the
colony. They stay mostly within the nest. On the warmest days a few
will come out and walk leisurely around the mound. They are not
scattered irregularly through the nest, but seem to be housed together
in large chambers. In one of these chambers I found a wingless queen in
their midst. It seemed very fitting for a queen to be surrounded by
Amazon soldiers; but, alas! they seemed more like maids of honor than
soldiers, for they forsook the royal lady without making an effort to
defend her. Not so, however, with the little workers: they rallied
around her, ready to guard her with their lives, and no doubt would have
succeeded had it been any ordinary foe.

This phenomenon--the soldiers and queens with smooth mandibles--is very
puzzling, and has excited much interest among naturalists both in this
country and in Europe. I sent specimens to Mr. Charles Darwin, which he
forwarded to Mr. Frederick Smith of the British Museum (who, Mr. Darwin
informs me, is the highest authority in Europe on ants and other
Hymenoptera). Mr. Smith says: "Your observations on the structural
differences in the mandibles of this ant are quite new to me." I also
sent specimens to the eminent naturalist Dr. Auguste Forel of Munich,
who, like Mr. Smith, had never observed this feature of the mandibles in
any ant; but he has a theory to account for it--that the smooth
mandibles have been worn down by labor. If this theory is true, how can
we account for the fact that other ants do not wear down their teeth?
The chitinous covering of this harvesting-ant is firm and hard. The
stage forceps of my microscope closes with a spring, and in studying
this ant I have put thousands of individuals to the test, holding them
in the forceps to examine their mandibles, and in no instance do I
recollect seeing one injured, while many other species are easily
injured by the forceps. Among these are the two large species of
carpenter-ant before mentioned, which work in stumps or f