Systematics and evolution

Systematics

Updated after Marples (1962), Acosta Hospitaleche (2004), and Ksepka et al. (2006). See the gallery for images of most living species.

ORDER SPHENISCIFORMES
Basal and unresolved taxa
Waimanu - basal (Middle-Late Paleocene)
Sphenisciformes gen. et sp. indet. CADIC P 21 (Leticia Middle Eocene of Punta Torcida, Argentina: Clarke et al. 2003)
Family Spheniscidae
Subfamily Palaeeudyptinae - Giant penguins (fossil)
Crossvallia (Cross Valley Late Paleocene of Seymour Island, Antarctica) - tentatively assigned to this subfamily
Anthropornis (Middle Eocene ?- Early Oligocene of Seymour Island, Antarctica)
Nordenskjoeld's Giant Penguin, Anthropornis nordenskjoeldi
Archaeospheniscus (Middle/Late Eocene - Late Oligocene)
Delphinornis (Middle/Late Eocene ?- Early Oligocene of Seymour Island, Antarctica) - tentatively assigned to this subfamily
Icadyptes (Late Eocene of Atacama Desert, Peru)
Marambiornis (Late Eocene -? Early Oligocene of Seymour Island, Antarctica) - tentatively assigned to this subfamily
Mesetaornis (Late Eocene -? Early Oligocene of Seymour Island, Antarctica) - tentatively assigned to this subfamily
Palaeeudyptes (Middle/Late Eocene - Late Oligocene)
Pachydyptes (Late Eocene)
Platydyptes (Late Oligocene of New Zealand)
Anthropodyptes (Middle Miocene) - tentatively assigned to this subfamily
Subfamily Paraptenodytinae - Patagonian stout-legged penguins (fossil)
Arthrodytes (San Julian Late Eocene/Early Oligocene - Patagonia Early Miocene of Patagonia, Argentina)
Paraptenodytes (Early - Late Miocene/Early Pliocene)
Subfamily Palaeospheniscinae - Patagonian slender-legged penguins (fossil)
Eretiscus (Patagonia Early Miocene of Patagonia, Argentina)
Palaeospheniscus (Early? - Late Miocene/Early Pliocene) - includes Chubutodyptes
Subfamily Spheniscinae - Modern penguins
Aptenodytes - Great penguins (2 species)
Pygoscelis - Brush-tailed penguins (3 species)
Eudyptula - Little penguins (2 species)
Spheniscus - Banded penguins (4 species)
Megadyptes - Yellow-eyed Penguin
Eudyptes - Crested penguins (5-7 living species, possibly another one recently extinct)
Not assigned to a subfamily (all fossil)
Perudyptes (Middle Eocene of Atacama Desert, Peru) - basal?
Tonniornis (Late Eocene -? Early Oligocene of Seymour Island, Antarctica)
Wimanornis (Late Eocene -? Early Oligocene of Seymour Island, Antarctica)
Duntroonornis (Late Oligocene of Otago, New Zealand)
Korora (Late Oligocene of S Canterbury, New Zealand)
Spheniscidae gen. et sp. indet (Late Oligocene/Early Miocene of Hakataramea, New Zealand)
Pseudaptenodytes (Late Miocene/Early Pliocene)
Marplesornis (Pliocene)
Dege - possibly Spheniscinae
Nucleornis - possibly Spheniscinae

Taxonomy: Clarke et al. (2003) and Ksepka et al. (2006) apply the phylogenetic taxon Spheniscidae to what here is referred to as Spheniscinae. Furthermore, they restrict the phylogenetic taxon Sphenisciformes to flightless taxa, and establish (Clarke et al. 2003) the phylogenetic taxon Pansphenisciformes as equivalent to the Linnean taxon Sphenisciformes, i.e., including any flying basal "proto-penguins" to be discovered eventually. Given that neither the relationships of the penguin subfamilies to each other nor the placement of the penguins in the avian phylogeny is presently resolved, this seems spurious and in any case is confusing; the established Linnean system is thus followed here.

Evolution

The evolutionary history of penguins is by now fairly well researched and represents a showcase of evolutionary biogeography; though as penguin bones of any one species vary much in size and few good specimens are known, the alpha taxonomy of many prehistoric forms still leaves much to be desired. A number of seminal articles dealing with penguin prehistory has been published since 2005 (Bertelli & Giannini 2005, Baker et al. 2006, Ksepka et al. 2006, Slack et al. 2006), and at least the evolution of the living genera can be considered resolved by now.

According to the comprehensive review of the available evidence by Ksepka et al. (2006), the basal penguins lived around the time of the Cretaceous-Paleogene extinction event somewhere in the general area of (southern) New Zealand and Byrd Land, Antarctica. Due to plate tectonics, these areas were at that time less than 1500 kilometers apart rather than the 4000 km of today. The last common ancestor of penguins and their sister clade can be roughly dated to the Campanian-Maastrichtian boundary, around 70-68 mya (Baker et al. 2006, Slack et al. 2006)[4] What can be said as certainly as possible in the absence of direct (i.e., fossil) evidence is that by the end of the Cretaceous, the penguin lineage must have been evolutionarily well distinct, though much less so morphologically; it is fairly likely that they were not yet entirely flightless at that time, as flightless birds have generally low resilience to the breakdown of trophic webs which follows the initial phase of mass extinctions because of their below-average dispersal capabilities (see also Flightless Cormorant).

The basal fossils

The oldest known fossil penguin species is Waimanu manneringi, which lived in the early Paleocene epoch of New Zealand, or about 62 mya (Slack et al. 2006). While they were not as well adapted to aquatic life as modern penguins, Waimanu were generally loon-like birds but already flightless, with short wings adapted for deep diving. They swam on the surface using mainly their feet, but the wings were - as opposed to most other diving birds living and extinct - already adapting to underwater locomotion.

An unnamed fossil from Argentina proves that by the Bartonian (Middle Eocene), some 39-38 mya[5], primitive penguins had spread to South America and were in the process of expanding into Atlantic waters (Clarke et al. 2003).

“Palaeëudyptines”

During the Late Eocene and the Early Oligocene (40-30 mya), some lineages of gigantic penguins existed. Nordenskjoeld's Giant Penguin was the tallest, growing nearly 1.80 meters (6 feet) tall. The New Zealand Giant Penguin was probably the heaviest, weighing 80 kg or more. Both were found on New Zealand, the former also in the Antarctic farther eastwards.

Traditionally, most extinct species of penguins, giant or small, had been placed in the paraphyletic subfamily called Palaeeudyptinae. More recently, with new taxa being discovered and placed in the phylogeny if possible, it is becoming accepted that there were at least 2 major extinct lineages. One or two closely related ones occurred in Patagonia, and at least one other - which is or includes the paleeeudyptines as recognized today - occurred on most Antarctic and subantarctic coasts.

But size plasticity seems to have been great at this initial stage of penguin radiation: on Seymour Island, Antarctica, for example, around ten known species of penguins ranging from medium to huge size apparently coexisted some 35 mya during the Priabonian (Late Eocene) (Jadwiszczak 2006). It is not even known whether the gigantic palaeeudyptines constitute a monophyletic lineage, or whether gigantism was evolved independently in a much restricted Palaeeudyptinae and the Anthropornithinae - were they considered valid -, or whether there was a wide size range present in the Palaeeudyptinae as delimited as usually done these days (i.e., including Anthropornis) (Ksepka et al. 2006). The oldest well-described giant penguin, the 5-foot-tall Icadyptes salasi, actually occurred as far north as northern Peru about 36 mya.

In any case, the gigantic penguins had disappeared by the end of the Paleogene, around 25 mya. Interestingly, their decline and disappearance coincides with the spread of the Squalodontoidea and other primitive, fish-eating toothed whales, which certainly competed with them for food, and were ultimately more successful (Baker et al. 2006). A new lineage, the Paraptenodytinae which includes smaller but decidedly stout-legged forms, had already arisen in southernmost South America by that time. The early Neogene saw the emergence of yet another morphotype in the same area, the similarly-sized but more gracile Palaeospheniscinae, as well as the radiation which gave rise to the penguin biodiversity of our time.

Origin and systematics of modern penguins

Modern penguins consititute two undisputed clades and another two more basal genera with more ambiguous relationships (Bertelli & Giannini 2005). The origin of the Spheniscinae lies probably in the latest Paleogene, and geographically it must have been much the same as the general area in which the order evolved: the oceans between the Australia-New Zealand region and the Antarctic (Baker et al. 2006). Presumedly diverging from other penguins around 40 mya (Baker et al. 2006), it seems that the Spheniscinae were for quite some time limited to their ancestral area, as the well-researched deposits of the Antarctic Peninsula and Patagonia have not yielded Paleogene fossils of the subfamily. Also, the earliest spheniscine lineages are those with the most southern distribution.

The genus Aptenodytes appears to be the basalmost divergence among living penguins; they have bright yellow-orange neck, breast, and bill patches, incubate by placing their eggs on their feet, and when they hatch, they are almost naked. This genus has a distribution centered on the Antarctic coasts and barely extends to some subantarctic islands today.

Pygoscelis contains species with a fairly simple black-and-white head pattern; their distribution is intermediate, centered on Antarctic coasts but extending somewhat northwards from there. In external morphology, these apparently still resemble the common ancestor of the Spheniscinae, as Aptenodytes' autapomorphies are in most cases fairly pronounced adaptations related to that genus' extreme habitat conditions. As the former genus, Pygoscelis seems to have diverged during the Bartonian[6], but the range expansion and radiation which lead to the present-day diversity probably did not occur until much later, around the Burdigalian stage of the Early Miocene, roughly 20-15 mya (Baker et al. 2006).

The genera Spheniscus and Eudyptula contain species with a mostly subantarctic distribution centered on South America; some, however, range quite far northwards. They all lack carotenoid coloration, and the former genus has a conspicuous banded head pattern; they are unique among living penguins in nesting in burrows. This group probably radiated eastwards with the Antarctic Circumpolar Current out of the ancestral range of modern penguins throughout the Chattian (Late Oligocene), starting approximately 28 mya (Baker et al. 2006). While the two genera separated during this time, the present-day diversity is the result of a Pliocene radiation, taking place some 4-2 mya (Baker et al. 2006).

The Megadyptes - Eudyptes clade occurs at similar latitudes (though not as far north as the Galapagos Penguin), has its highest diversity in the New Zealand region, and represent a westward dispersal. They are characterized by hairy yellow ornamental head feathers; their bills are at least partly red. These two genera diverged apparently in the Middle Miocene (Langhian, roughly 15-14 mya), but again, the living species of Eudyptes are the product of a later radiation, stretching from about the late Tortonian (Late Miocene, 8 mya) to the end of the Pliocene (Baker et al. 2006).

It is most interesting to note that the geographical and temporal pattern or spheniscine evolution corresponds closely to two episodes of global cooling documented in the paleoclimatic record (Baker et al. 2006). The emergence of the subantarctic lineage at the end of the Bartonian corresponds with the onset of the slow period of cooling that eventually led to the ice ages some 35 million years later. With habitat on the Antarctic coasts declining, by the Priabonian more hospitable conditions for most penguins existed in the subantarctic regions rather than in Antarctica itself. Notably, the cold Antarctic Circumpolar Current also started as a continuous circumpolar flow only around 30 mya, on the one hand forcing the Antarctic cooling, and on the other faciliating the eastward expansion of Spheniscus to South America and eventually beyond (Baker et al. 2006).

Later, an interspersed period of slight warming was ended by the Middle Miocene Climate Transition, a sharp drop in global average temperature from 14 to 12 mya, and similar abrupt cooling events followed at 8 mya and 4 mya; by the end of the Tortonian, the Antarctic ice sheet was already much like today in volume and extent. The emergence of most of today's subantarctic penguin species almost certainly was caused by this sequence of Neogene climate shifts.

Relationship to other bird orders

Penguin ancestry beyond Waimanu remains unknown and not well resolved by molecular or morphological analyses. The latter tend to be confounded by the strong adative autapomorphies of the Sphenisciformes; a sometimes perceived fairly close relationship between penguins and grebes is almost certainly an error based on both groups' strong diving adaptations, which are homoplasies. On the other hand, different DNA sequence datasets do not agree in detail with each other either.

What seems clear is that penguins belong to a clade of Neoaves (living birds except paleognaths and fowl) which comprises of what is sometimes called "higher waterbirds" to distinguish them from the more ancient waterfowl. This group contains such birds as storks, rails, and the seabirds, with the possible exception of the Charadriiformes (Fain & Houde 2004).

Inside this group, penguin relationships are far less clear. Depending on the analysis and dataset, a close relationship to Ciconiiformes (e.g. Slack et al. 2006) or to Procellariiformes (Baker et al. 2006) has been suggested. Some (e.g. Mayr 2005) think the penguin-like plotopterids (usually considered relatives of anhingas and cormorants) may actually be a sister group of the penguins, and that penguins may have ultimately shared a common ancestor with the Pelecaniformes and consequently would have to be included in that order, or that the plotopterids were not as close to other pelecaniforms as generally assumed, which would necessitate splitting the traditional Pelecaniformes in three.

The Auk of the Northern Hemisphere is superficially similar to penguins, they are not related to the penguins at all, but considered by some to be a product of moderate convergent evolution [7]