Submerged swimming of the great cormorant Phalacrocorax carbo sinensis is a variant of the burst-and-glide gait.

Cormorants are water birds that forage by submerged swimming in search and pursuit of fish. Underwater they swim by paddling with both feet simultaneously in a gait that includes long glides between consecutive strokes. At shallow swimming depths the birds are highly buoyant as a consequence of their aerial lifestyle. To counter this buoyancy cormorants swim underwater with their body at an angle to the swimming direction. This mechanical solution for foraging at shallow depth is expected to increase the cost of swimming by increasing the drag of the birds. We used kinematic analysis of video sequences of cormorants swimming underwater at shallow depth in a controlled research setup to analyze the swimming gait and estimate the resultant drag of the birds during the entire paddling cycle. The gliding drag of the birds was estimated from swimming speed deceleration during the glide stage while the drag during active paddling was estimated using a mathematical ;burst-and-glide' model. The model was originally developed to estimate the energetic saving from combining glides with burst swimming and we used this fact to test whether the paddling gait of cormorants has similar advantages. We found that swimming speed was correlated with paddling frequency (r=0.56, P<0.001, N=95) where the increase in paddling frequency was achieved mainly by shortening the glide stage (r=-0.86, P<0.001, N=95). The drag coefficient of the birds during paddling was higher on average by two- to threefold than during gliding. However, the magnitude of the drag coefficient during the glide was positively correlated with the tilt of the body (r=0.5, P<0.003, N=35) and negatively correlated with swimming speed (r=-0.65, P<0.001, N=35), while the drag coefficient during the stroke was not correlated with tilt of the body (r=-0.11, P>0.5, N=35) and was positively correlated with swimming speed (r=0.41, P<0.015, N=35). Therefore, the difference between the drag coefficient during the glide and during propulsion diminished at lower speeds and larger tilt. The mean drag of the birds for a single paddling cycle at an average swimming speed of 1.5 m s(-1) was 5.5+/-0.68 N. The burst-and-glide model predicts that energy saving from using burst-and-glide in the paddling cycle is limited to relatively fast swimming speeds (>1.5 m s(-1)), but that as the birds dive deeper (>1 m where buoyancy is reduced), the burst-and-glide gait may become beneficial even at lower speeds.