Loggers were attached to free-ranging Brünnich's guillemots Uria lomvia during dives, to measure swim speeds, body angles, stroke rates, stroke and glide durations, and acceleration patterns within strokes, and the data were used to model the mechanical costs of propelling the body fuselage (head and trunk excluding wings). During vertical dives to 102-135·m, guillemots regulated their speed during descent and much of ascent to about 1.6±0.2·m·s -1 . Stroke rate declined very gradually with depth, with little or no gliding between strokes. Entire strokes from 2·m to 20·m depth had similar forward thrust on upstroke vs downstroke, whereas at deeper depths and during horizontal swimming there was much greater thrust on the downstroke. Despite this distinct transition, these differences had small effect (<6%) on our estimates of mechanical cost to propel the body fuselage, which did not include drag of the wings. Work·stroke -1 was quite high as speed increased dramatically in the first 5·m of descent against high buoyancy. Thereafter, speed and associated drag increased gradually as buoyancy slowly declined, so that mechanical work·stroke -1 during the rest of descent stayed relatively constant. Similar work·stroke -1 was maintained during non-pursuit swimming at the bottom, and during powered ascent to the depth of neutral buoyancy (about 71·m). Even with adjustments in respiratory air volume of ±60%, modeled work against buoyancy was important mainly in the top 15·m of descent, after which almost all work was against drag. Drag was in fact underestimated, as our values did not include enhancement of drag by altered flow around active swimmers. With increasing buoyancy during ascent above 71·m, stroke rate, glide periods, stroke acceleration patterns, body angle and work·stroke -1 were far more variable than during descent; however, mean speed remained fairly constant until buoyancy increased rapidly near the surface. For dives to depths >20·m, drag is by far the main component of mechanical work for these diving birds, and speed may be regulated to keep work against drag within a relatively narrow range.

Summary1. For birds, mammals and turtles, costs of swimming by foot propulsion are usually much higher than for propulsion by wings or foreflippers. The propulsive efficiency with which limbs impart thrust to the water is greater for lift-based wings than for drag-based feet, but different acceleration patterns during oscillatory strokes may also alter total drag on the body fuselage (head and trunk). 2. Because wing propulsion allows thrust on both upstroke and downstroke, whereas foot propulsion in many species (perhaps excepting grebes) has little or no thrust on the upstroke, foot propulsion requires higher speeds during a smaller fraction of the stroke to maintain the same mean speed. Because drag increases non-linearly with increasing speed, higher instantaneous speeds in drag-based foot propulsion may cause greater total drag on the body fuselage. 3. Tow-tank measurements have shown that foot-propelled birds that swim with long necks extended have lower fuselage drag at high speeds than do wing-propelled birds that swim with necks retracted. This difference might reduce the higher costs of drag-based foot propulsion, but such effects must be evaluated in the context of drag at a range of speeds throughout oscillatory strokes. 4. In quasi-steady models of horizontal swimming underwater, stroke-acceleration curves for both foot and wing propulsion were applied to a range of bird shapes and sizes. Higher fuselage drag during foot propulsion increased mechanical costs of transport (MCOT, J kg -1 m -1 ) by 26-40% in various species. Thus, a large fraction of the different costs of wing and foot propulsion might be explained in terms of drag on the body fuselage, independently of the propulsive efficiency of stroking limbs. 5. When drag curves for different body shapes were combined with different oscillatory stroking patterns, swimming with a long neck extended did not reduce the higher total drag associated with drag-based foot propulsion. Thus, although size and shape can affect drag measured at different constant speeds, effects of drag on locomotor costs depend much more on stroke-acceleration patterns of different swimming modes.

For birds diving to depths where pressure has mostly reduced the buoyancy of air spaces, hydrodynamic drag is the main mechanical cost of steady swimming. Drag is strongly affected by body size and shape, so such differences among species should affect energy costs. Because flow around the body is complicated by the roughness and vibration of feathers, feathers must be considered in evaluating the effects of size and shape on drag. We investigated the effects of size, shape and feathers on the drag of avian divers ranging from wing-propelled auklets weighing 75 g to foot-propelled eiders weighing up to 2060 g. Laser scanning of body surfaces yielded digitized shapes that were averaged over several specimens per species and then used by a milling machine to cut foam models. These models were fitted with casts of the bill area, and their drag was compared with that of frozen specimens. Because of the roughness and vibration of the feathers, the drag of the frozen birds was 2–6 times that of the models. Plots of drag coefficient (C(D)) versus Reynolds number (Re) differed between the model and the frozen birds, with the pattern of difference varying with body shape. Thus, the drag of cast models or similar featherless shapes can differ both quantitatively and qualitatively from that of real birds. On the basis of a new towing method with no posts or stings that alter flow or angles of attack, the dimensionless C(D)/Re curves differed among a size gradient of five auklet species (75–100g) with similar shapes. Thus, extrapolation of C(D)/Re curves among related species must be performed with caution. At lower speeds, the C(D) at a given Re was generally higher for long-necked birds that swim with their neck extended (cormorants, grebes, some ducks) than for birds that swim with their head retracted (penguins, alcids), but this trend was reversed at high speeds. Because swimming birds actually travel at a range of instantaneous speeds during oscillatory strokes, species variations in drag at different speeds must be considered in the context of accelerational stroking.