For my birthday, my parents sent DVDs of the Life series with David Attenborough, nature films from the BBC. These are wonderfully made, with amazing photography, and you will likely never hear me say that about any other films. One aspect of the photography that is amazing are the underwater shots. This has got me thinking about marine mammals and seabirds, and the differences between them.
There are five independent lineages of extant (extant is the opposite of extinct) marine mammals:
1. The Cetaceans (whales, porpoises and dolphins), relatives of pigs and hippos.
2. The Sirenians (manatees and dugongs), mildly related to elephants and hyraxes.
3/4. The Pinnepeds (walrus, seals and sea lions), descended from a dog-like carnivore.
3/4. The sea otter, an otter, which is an aquatic weasel.
5. The polar bear, bear.
I have ordered these in the degree to which they have become fully aquatic. The Cetaceans neither need to, nor safely can, leave the water. This is true of Sirenians also, but they tend to feed and birth in shallow water near shorelines, where whales wander the open oceans and dive to amazing depths. The Pinnepeds aren't so good on land, but they do haul up to breed and pup. The sea otter is in some ways more fully marine than the Pinnepeds, mating and usually giving birth at sea. But again, otters are more tied to the land than are Pinnepeds the rest of the year, living and feeding in coastal kelp forests and being capable of fast and efficient movement on land. The polar bear is marine in that it swims long distances, hunts at sea, and has structures that specifically help it do these things. But it still prefers to walk rather than swim, brings its food onto solid ground to feed, breeds and pups out of the water and so forth.
There are more independent groups of sea birds, even if you don't consider each transition from freshwater to saltwater. Penguins are perhaps the most fully marine, flying only in water, feeding entirely on seafood, having special mechanisms for dealing with high levels of salt. The Procellariiformes (albatrosses and petrels) are not far behind, spending about as much time at sea (although over rather than in) as penguins do. The Phaethontiformes (tropicbirds) spend most of their lives at sea, as do many of the Pelicaniformes (pelicans, frigatebirds, boobies, gannets, cormorants and shags). While many gulls live far from the sea, many Charadriiformes (gulls, terns, skuas, plovers, puffins, auks etc.) are extremely marine. Many Anseriformes (ducks, swans and geese), particularly the Merginae (sea ducks) are, well, sea ducks. It was recently discovered that gyrfalcons spend long periods hunting on and around sea-ice, although they probably don't actually swim. I'm sure I've forgotten other examples.
Why so many birds moving out to sea, but so few mammals?
The obvious first hypothesis is that the ability to fly is very useful at sea, while the ability to walk/run/hop etc. is not. The falcons are a pretty terrestrial group, but with few changes beyond the behavioral, gyrfalcons can spend extended periods at sea. Ospreys and eagles, relatives of falcons, use talons that evolved grabbing terrestrial prey to scoop fish. Even hummingbirds and warblers that can't forage or land at sea regularly spend long periods migrating over open ocean. Birds may have an easier entree than do mammals.
Given this, it may be surprising that the most fully marine descendent of terrestrial vertebrates are not birds. All birds lay eggs, and none have figured out how to make that work at sea, so all need to maintain the ability to be land animals. Almost all mammals give live birth, and three groups (Cetaceans, Sirineans and sea otters) can do that without ever leaving the water. The birds may have an easier time getting started down evolutionary paths that lead to a marine life, but they seem to have an inescapable constraint that keeps them from finishing that path: shelled eggs.
The marine reptiles show an interesting parallel to this. Marine iguanas, saltwater crocs and sea turtles all lay eggs, and all do so on land. Sea snakes, excepting one genus, birth live young, and do so at sea. That one genus lays eggs on land.
If some snakes have evolved the ability to have their eggs hatch internally and their hatchlings ready to swim the moment they emerge from the mother, why can't some bird do the same? Imagine how much better off an emperor penguin would be if instead of spending the Antarctic winter fasting in the cold, it could spend that time feasting in the ocean with it's chick developing internally.
Any answer I could offer would be pure speculation. One class of question that evolutionary biology is very bad at answering is "why didn't X evolve." Why hasn't any bird evolved live birth? Maybe it is something about their egg shells. Maybe they are in a habitat where that just doesn't work. Probably it just never happened.