The arena of a large bottle enables each athlete to keep out of the other’s way, respected by her adversary, who is respected in her turn.  Let us reduce the lists, diminish the enclosure.  I put Bumble-bee and Tarantula into a test-tube that has only room for one at the bottom.  A lively brawl ensues, without serious results.  If the Bumble-bee be underneath, she lies down on her back and with her legs wards off the other as much as she can.  I do not see her draw her sting.  The Spider, meanwhile, embracing the whole circumference of the enclosure with her long legs, hoists herself a little upon the slippery surface and removes herself as far as possible from her adversary.  There, motionless, she awaits events, which are soon disturbed by the fussy Bumble-bee.  Should the latter occupy the upper position, the Tarantula protects herself by drawing up her legs, which keep the enemy at a distance.  In short, save for sharp scuffles when the two champions are in touch, nothing happens that deserves attention.  There is no duel to the death in the narrow arena of the test-tube, any more than in the wider lists afforded by the bottle.  Utterly timid once she is away from home, the Spider obstinately refuses the battle; nor will the Bumble-bee, giddy though she be, think of striking the first blow.  I abandon experiments in my study.

We must go direct to the spot and force the duel upon the Tarantula, who is full of pluck in her own stronghold.  Only, instead of the Bumble-bee, who enters the burrow and conceals her death from our eyes, it is necessary to substitute another adversary, less inclined to penetrate underground.  There abounds in the garden, at this moment, on the flowers of the common clary, one of the largest and most powerful Bees that haunt my district, the Carpenter-bee (Xylocopa violacea), clad in black velvet, with wings of purple gauze.  Her size, which is nearly an inch, exceeds that of the Bumble-bee.  Her sting is excruciating and produces a swelling that long continues painful.  I have very exact memories on this subject, memories that have cost me dear.  Here indeed is an antagonist worthy of the Tarantula, if I succeed in inducing the Spider to accept her.  I place a certain number, one by one, in bottles small in capacity, but having a wide neck capable of surrounding the entrance to the burrow.

As the prey which I am about to offer is capable of overawing the huntress, I select from among the Tarantulae the lustiest, the boldest, those most stimulated by hunger.  The spikeleted stalk is pushed into the burrow.  When the Spider hastens up at once, when she is of a good size, when she climbs boldly to the aperture of her dwelling, she is admitted to the tourney; otherwise, she is refused.  The bottle, baited with a Carpenter-bee, is placed upside down over the door of one of the elect.  The Bee buzzes gravely in her glass bell; the huntress mounts from the recesses of the cave; she is on the threshold, but inside; she looks; she waits.  I also wait.  The quarters, the half-hours pass:  nothing.  The Spider goes down again:  she has probably judged the attempt too dangerous.  I move to a second, a third, a fourth burrow:  still nothing; the huntress refuses to leave her lair.