Our knowledge of the form of lateralized sleep behavior, known as unihemispheric slow wave sleep (USWS), seen in all members of the order Cetacea examined to date, is described. We trace the discovery of this phenotypically unusual form of mammalian sleep and highlight specific aspects that are different from sleep in terrestrial mammals. We find that for cetaceans sleep is characterized by USWS, a negligible amount or complete absence of rapid eye movement (REM) sleep, and a varying degree of movement during sleep associated with body size, and an asymmetrical eye state. We then compare the anatomy of the mammalian somnogenic system with what is known in cetaceans, highlighting areas where additional knowledge is needed to understand cetacean sleep. Three suggested functions of USWS (facilitation of movement, more efficient sensory processing and control of breathing) are discussed. Lastly, the possible selection pressures leading to this form of sleep are examined, leading us to the suggestion that the selection pressure necessitating the evolution of cetacean sleep was most likely the need to offset heat loss to the water from birth and throughout life. Aspects such as sentinel functions and breathing are likely to be proximate evolutionary phenomenon of this form of sleep.

A bottlenose porpoise was trained to dive untethered in the open ocean and to exhale into an underwater collecting funnel before surfacing from prescribed depths down to 300 meters. The animal was also taught to hold its breath for periods up to 4 minutes at the surface and then blow in the funnel. Alveolar collapse is probably complete at around 100 meters, and little pulmonary respiratory exchange occurs below that depth. Thoracic collapse was observ visually at 10 to 50 meters and by underwater television to 300 meters.

Intramuscular nitrogen tensions in Tursiops truncatus after a schedule of repetitive ocean dives suggest a lung collapse depth of about 70 meters and suggest that intramuscular circulation is maintained during unrestrained diving in the open ocean. Therefore, the bottle-nosed dolphin is not protected by lung collapse from the decompression hazards of dives to depths shallower than 70 meters.

Small toothed whales of the family Phocoenidae and delphinid genus Cephalorhynchus produce long-duration, narrowband biosonar clicks above 100 kHz, that might be seen as an adaptation for short range echolocation in shallow water. This paper presents data showing that the distantly related, and larger pygmy sperm whale Kogia breviceps (Kogiidae), that is a deep-diving, cephalopod-eating toothed whale, produce narrow-banded high frequency (NBHF) clicks identical to those of Phocoena and Cephalorhynchus (f 0 = 130kHz, Q 3 ds>10, duration > 80 msec). Thus, NBHF biosonar signals have evolved on three independent occasions in the odontocete suborder, but the apparent functional convergence does not relate to anatomical or niche similarity. Rather, it is suggested that a biosonar strategy adapting to a minimum in ocean ambient noise above 100 kHz in concert with high Q auditory filters have led to convergent evolution of the NBHF biosonar clicks. Since these biosonar signals carry all their energy at frequencies above the upper hearing limit of the killer whale Orcinus orca, predator avoidance may also have been a evolutionary shaping factor of the sonar signals from these non-whistling odontocetes.

We compared mature dolphins with 4 other groupings of mature cetaceans. With a large data set, we found great brain diversity among 5 different taxonomic groupings. The dolphins in our data set ranged in body mass from about 40 to 6,750 kg and in brain mass from 0.4 to 9.3 kg. Dolphin body length ranged from 1.3 to 7.6 m. In our combined data set from the 4 other groups of cetaceans, body mass ranged from about 20 to 120,000 kg and brain mass from about 0.2 to 9.2 kg, while body length varied from 1.21 to 26.8 m. Not all cetaceans have large brains relative to their body size. A few dolphins near human body size have human-sized brains. On the other hand, the absolute brain mass of some other cetaceans is only one-sixth as large. We found that brain volume relative to body mass decreases from Delphinidae to a group of Phocoenidae and Monodontidae, to a group of other odontocetes, to Balaenopteroidea, and finally to Balaenidae. We also found the same general trend when we compared brain volume relative to body length, except that the Delphinidae and Phocoenidae-Monodontidae groups do not differ significantly. The Balaenidae have the smallest relative brain mass and the lowest cerebral cortex surface area. Brain parts also vary. Relative to body mass and to body length, dolphins also have the largest cerebellums. Cortex surface area is isometric with brain size when we exclude the Balaenidae. Our data show that the brains of Balaenidae are less convoluted than those of the other cetaceans measured. Large vascular networks inside the cranial vault may help to maintain brain temperature, and these nonbrain tissues increase in volume with body mass and with body length ranging from 8 to 65% of the endocranial volume. Because endocranial vascular networks and other adnexa, such as the tentorium cerebelli, vary so much in different species, brain size measures from endocasts of some extinct cetaceans may be overestimates. Our regression of body length on endocranial adnexa might be used for better estimates of brain volume from endocasts or from endocranial volume of living species or extinct cetaceans.