1. 1.

    Analysis of cine films and intramantle pressure records for squid Loligo opalescens Berry swimming in a tunnel respirometer provided estimates of all the forces acting in the horizontal and vertical planes for swimming speeds from 0.1 to 0.5 ms−1.

  2. 2.

    Different speeds used different gaits; fin thrust was only important below 0.2 ms−1, ‘anaerobic’ circular muscles were recruited only at supracritical speeds, and hyperinflation caused by contraction of the radial muscle was not seen in steady swimming.

  3. 3.

    The extent, rate and frequency of contraction of the obliquely striated circular muscles varied little with speed, and jet thrust was matched to speed primarily by active pressure control through adjustments in the size of the funnel orifice.

  4. 4.

    Hydrodynamic lift production to compensate for negative buoyancy during enforced horizontal swimming in the tunnel required 30–90% of the total force over the speed range studied and appears less efficient than direct use of jet thrust. This suggests a new rationale for ‘climb-and-glide’ swimming which reduces previous estimates of the gross cost of transport for squid under natural conditions by at least 35%, with no loss of speed.

  5. 5.

    The cost of accelerating water into the mantle of a squid moving at high speed appears to have been underestimated in previous studies. A simulation of a series of escape jets predicts a maximum speed of 8 body lengths s−1 (1.4ms−1), reached after only two jets, because of the high deceleration during refilling.

This content is only available via PDF.