The work at the dam site began at mid-morning and continued for one hour and forty minutes. Zephyr directed operations from a position three wingspans above the dam's surface, hovering with the precise, energy-expensive stability that only dragonflies among the insect world could maintain for extended periods.
The water striders operated in shifts on the downstream water face of the obstruction. They moved in coordinated groups of six, creating synchronized wave patterns by raising and lowering their leg positions in sequence — an action that individually produced trivial water movement but collectively created a resonant pressure wave that concentrated force at the base of the loose clay section.
Zephyr could observe the clay face responding to this pressure over the first thirty minutes as a series of small cracks propagated upward from the water line.
The diving beetles' work was slower and more physically demanding. Each beetle could carry a stone no larger than a pea in its leg grip, but with eleven beetles working the gravel transfer operation in rotation, the redirecting structure grew steadily along the channel floor.
The structure they were building was a simple crescent-shaped arrangement of stones that would force the current to angle directly into the base of the target section rather than flowing around it. Zephyr watched its progress through the distorting lens of the water surface, reading the changing turbulence patterns above the construction site as each stone was placed.
His own contribution was the most technically demanding element. The wing-vortex technique required Zephyr to beat his wings in a specific asymmetric pattern that generated a downward-spiraling air column rather than the normal lift-generating stroke. Sustained correctly, this produced a concentrated pressure pulse on the surface below him — not enough to move clay, but enough to drive air into existing cracks, expanding them fractionally with each pulse cycle.
He sustained the vortex generation for forty-five consecutive minutes, pausing for one-minute rest intervals every fifteen minutes. The technique was exhausting in a way that normal flight was not, demanding fine motor control throughout his entire wing musculature while maintaining the hovering position against the upstream breeze.
The breakthrough came at the one-hour-twenty-minute mark. A section of the loose clay mass approximately thirty centimeters wide detached from the lower face of the dam and was immediately taken by the current, breaking into smaller pieces as it moved downstream.
The gap it left was not large — perhaps twenty centimeters of clear passage at the base of the center section — but the current entering through the gap immediately began working on the exposed clay faces on either side, widening the opening progressively.
Within twenty minutes, the gap had grown to forty centimeters and the upstream water level had visibly dropped. The silt concentration in the downstream channel began decreasing as the total clay mass of the dam reduced.
The physical demands of sustained vortex generation drew on energy reserves that Zephyr had not depleted at this rate since the previous season's flood response operation. Dragonfly flight metabolism operates at an extremely high rate — their oxygen consumption per unit of body mass during active flight exceeds that of any other flying animal, including birds and bats.
The asymmetric vortex generation required for pressure-pulse work was approximately forty percent more energy-intensive than normal cruising flight.
He ate twice during the operation — brief predatory darts from his station above the dam to intercept the small midges that the silt-laden water was driving to the surface in unusual numbers. Each feeding stop lasted less than three seconds: departure from hover, interception, return to station, wing-beat pattern resumption.
The energy recovered from two midges was sufficient to sustain the operation for another twenty minutes. Zephyr incorporated these feeding stops into the rest-cycle intervals, maintaining the pressure-pulse pattern for fourteen minutes and then resting and feeding during the one-minute recovery period.
The diving beetles worked without rest breaks. Their oxygen storage system — the air bubble beneath the wing covers — allowed them to maintain active work at depth for periods of up to forty-five seconds before needing to surface.
The gravel transfer operation they were conducting required approximately fifteen-second work intervals per stone, which meant each beetle could place three stones per dive with a comfortable oxygen margin. Over one hour and twenty minutes with eleven beetles working, the current-redirecting structure grew to a total of approximately two hundred individual stones, all precisely positioned along the channel floor.
The water striders worked in rotation, with eighteen-individual work shifts alternating with twenty-individual resting groups. The wave-generation effort required each working strider to maintain a specific leg-position cycling rhythm that was physically taxing over extended periods. The rotation system allowed the full colony to sustain the effort without individual exhaustion.