- See also: Phylogeny
To succeed in spec evo, an important thing to know is how different groups of organisms relate to each other. Often, this can be quite confusing, so this page is a rough guide to phylogeny. This focuses on animals, as other kingdoms are not my expertise. However, anyone with expertise on plants, fungi, bacteria, or other kingdoms may edit this to include these kingdoms.
First, here are a few pointers when creating/using taxon names. Try to make all taxa clades, unless this is completely impossible, such as an entire order of animals stemming from a single species, but mention that the common ancestor of this group is that species, thus calling it a clade. Don't use paraphyletic groupings (such as Pelycosauria, lizards, monkeys, invertebrates, etc.) or polyphyletic ones (such as algae, Protozoa, worms, anapsids, etc.). When naming any clade with a rank from superfamily to subtribe in the case of animals (in some cases infratribe, but this is not officially recognized; infrafamily is also unrecognized), use correct terminations. These are -oidea, -idae, -inae, -ini, and -ina, for superfamily, family, subfamily, tribe, and subtribe, respectively. Also, remember to use Latin or Greek roots when giving formal names to species or clades. Also note that "taxon" is an alternative/favourable word over "genus" and "species" under phylogeny/cladistics. Now, with that out of the way, here is a guide to phylogeny.
Above the Kingdom Level
Starting at the top level, the original split of organisms occurred between bacteria and the all other organisms, the Archaea included. The next split was between the Archaea and the eukaryotes. Animals and fungi are within the Unikonta, and within that clade the Opisthokonta. In these clades there are also various protozoans, which are a polyphyletic group. Besides the Unikonta, the other large eukaryote groups are the Amoebozoa and the Bikonta, which includes the Protophyta and algae (both polyphyletic groups), as well as the plants.
Although this is not completely proven, the most basal animal phylum is likely to be the Porifera, or sponges. Their bodies are not separated into tissues, and the adults are completely sessile. Other, more radical, hypotheses include Ctenophora being the most basal phylum, but they share so many characteristics with the Cnidaria that it is unlikely that these features evolved independently. Also, Porifera are known from the fossil record from up to 760 million years ago (before either glaciation of the Cryogenian). Some scientists considered these finds to be dubious, but even if they are not sponges they are known from at least the Ediacaran.
After the rest of the animals diverged from the Porifera, they diverged into two main lineages, the Bilateria and Radiata. The Bilateria includes most living animals, and the Radiata includes radially symmetric animals. Recently, Radiata has been revised to be a clade, and now excludes the Porifera and Echinodermata, leaving only the Cnidaria and Ctenophora. Within the Bilateria, there are two main groups, the Deuterostomia, and the much more diverse Protostomia. In the Radiata, the two main groups are the Ctenophora, or comb jellies; and the Cnidaria, including jellyfish, sea anemones, and corals.
The Protostomia is the most diverse group of animals above the phylum level. It can be divided into three superphyla, the Ecdysozoa, the Lophotrochozoa, and the Platyzoa. The Ecdysozoa includes the familiar arthropods, the somewhat well known tardigrades and nematodes, and several lesser known phyla. The Lophotrochozoa is composed of the familiar molluscs and annelids, and other lesser known phyla, such as the brachiopods and the bryozoans. The Platyzoa contains the familiar flatworms, or Platyhelmithes, and other unfamiliar phyla, such as the microscopic rotifers.
Although less diverse than the protostomes, the deuterostomes include the vertebrates, likely the most successful group in animal history. Besides the vertebrates and other chordates, the Deuterostoma includes the Echinodermata and Hemichordata. Since these last two phyla do not have complicated evolutionary relationships, they will be detailed here.
Besides several extinct groups, the Echinodermata includes three main extant groups – the Crinozoa, the Asterozoa, and the Echinozoa. The Crinozoa includes the sessile crinoids (Crinoidea), and several other extinct classes. The Asterozoa includes the sea stars (Asteroidea) and the brittle stars (Ophiuroidea). In turn, the Ophiuroidea includes two main groups, the true brittle stars and the larger basket stars. The Echinozoa includes the sea urchins and sand dollars (Echinoidea), as well as the sea cucumbers (Holothuroidea).
The Hemichordata are considered the sister taxon to the Echinodermata, and have simple evolutionary relationships. They are composed of the classes Enteropneusta (acorn worms) and the Pterobranchia. They are generally considered to be a type of marine worm, although the marine worm group is polyphyletic and is a useless taxon.
Traditionally, the annelids were divided into the polychaetes, the oligochaetes, and leeches, but two of these groups proved to be paraphyletic. The Polychaeta gave rise to the other annelid groups, so they were paraphyletic. They are now divided into monophyletic groups, including the Myzostomida and the Echiura. The Oligochaeta, which include earthworms, gave rise to leeches (Hirudinea), so they are now combined in the class Citellata, referring to the reproductive organ shared by these hermaphrodites.
The arthropods are the most diverse phylum of animals, represented by well over a million known species, and millions yet to be discovered. All arthropods possess the trait of an exoskeleton, but like with vertebrates the defining structure is really all they have in common. They can be divided into five subphyla, which are the Chelicerata, Crustacea, Hexapoda, Myriapoda, and the extinct Trilobitomorpha. Each of these subphyla can in turn be divided into multiple classes.
The Triblobitamorphs are diverse grouping of extinct arthropods, most of which fall under the Class Trilobita. First appearing in the Cambrian, they greatly diminished in the Silurian and Devonian, before the final group was destroyed by the Permian-Triassic mass extinction. Their inner phylogeny is not perfectly understood, but their are generally eleven orders recognised, nine of which appeared in the Cambrian, the others right after it. This means they diversified very quickly, or have a ghost lineage. All but one order probably came from the paraphyletic suborder of Redlichiida, Redlichiina. Their outer relations are also unclear, but they are either basal matabulates or Myriochelates, depending on your favoured hypothoses. The Proetida were the last existing order of trilobites, going extinct at the end of the Palaeozoic.
The Chelicerata consists of three classes of arthropods, the Arachinda, Pycnogonida, and Xiphosura (also called Merostomata). The arachnids are the most common, and include many familiar groups, especially spiders (Araneae), scorpions (Scorpiones), and the ticks and mites (Acari). The Pycnogonida is an ancient class (and order, as it is monotypic) of aquatic arthropods, commonly called sea spiders. The last group, Xiphosura, (also called Merostomata) includes the modern horseshoe crabs, and the extinct eurypterids, or sea scorpions. However, the Xiphosura has been shown to be paraphyletic, as the smallest clade encompassing the horseshoe crabs and the eurypterids also encompasses the arachnids.
The Crustacea is a mostly aquatic (with a few exceptions, such as woodlice) subphylum of arthropods. Recently, its monophyly has been challenged, in the Pancrustacea hypothesis. It states that hexapods are actually derived crustaceans. Although there is almost no doubt that crustaceans are close relatives of the Hexapoda, hexapods being part of Crustacea is not definite, as the closest relatives of the Hexapoda in the hypothesis are the much older Branchiopoda. Crustaceans are divided into several classes, the most important of which is the Malacostraca, including decapods (crabs, lobsters, true shrimp, etc.), the isopods (several marine groups, and the terrestrial woodlice), and many other crustaceans. Another class is the Maxillopoda, but not only is it polyphyletic, no characteristic unites these groups. This includes crustaceans such as barnacles (Cirripedia) and copepods (Copepoda).
The Hexapoda are composed of two classes, the primitive Entognatha and the more advanced Insecta. Within the insects, there is a clade of flightless insects, the Apterygota, and a clade of mainly flying insects, the Pterygota. Many members of the Pterygota have secondarily lost flight, such as fleas, but they show similarities with other members of the Pterygota. The Pterygota is comprised of the Palaeoptera, including dragonflies and mayflies; and the Neoptera, which includes most living insects. Their internal relations have been and are still going over numerous revisions, such as the revalation of paraphyletic Mecoptera, which gave rise to Siphonaptera and Diptera. For more information on insect relationships, see here.
The Myriapoda are an arthropod group possessing the characteristic of large numbers of legs, although a few species possess as few as ten. They are likely the only group of animals to have a variable number of appendages. All myriapods are terrestrial. They are composed of four classes – the centipedes (Chilopoda), the millipedes (Diplopoda), the Paropoda (closely relating to the millipedes), and the Symphyla (sometimes called garden centipedes). Over 13,000 species of myriapods have been described. There are conflicting views on their placement in athropoda, with three main ideas, One goes by the classic crustacea hypothoses, and has them as sister to hexapods. The next follows pancrustacea and has it sister the that, forming Mandibulata. The last has the group on the other side of the tree, sister to Chelicerata, making up Myriochelata.
Now we reach the bulk of this page, the phylogeny of the Chordata. Starting at the top level, the chordates are composed of three subphyla – the Tunicata (tunicates, also known as Urochordata), Cephalochordata (lancelets), and Vertebrata (vertebrates). The relationships among these groups are unresolved, and two hypotheses have been presented. The first one, proposing a clade called Euchordata or Notochordata, groups together lancelets and vertebrates. An alternative clade, Olfactores, groups together tunicates and vertebrates.
The living cephalochordates are members of the order Amphioxiformes, inside the monotypic class Leptocardii. Lancelets lack eyes, and their "skin" is transparent. Because of their Agnathan-like appearance, they are sometimes considered the sister clades to chordates. Most living lancelets are members of the genus Branchiostoma. Two other living genera exist, Asymmetron and Epigonichthys. Fossil cephalochordates have been identified, the most famous of which is Pikaia, one of the most ancient known chordates.
Many more species of tunicates are known than lancelets, and tunicates comprise three classes. The first class, Ascidiacea, contains mostly sessile animals, known as sea squirts. The filter-feeding Thaliacea comprises mostly motile animals. The Larvacea live in the pelagic zone, and like the Thaliacea are motile. These three classes comprise over 2,000 known species of chordate.
The vertebrates are the dominant animals on Earth, and one of the most diverse groups, behind the arthropods and molluscs. All vertebrates have a spinal chord, except for the hagfish, which retain the primitive notochord. Vertebrates are divided into two clades, the Agnatha (jawless fish) and the Gnathostomata (jawed vertebrates). The Gnathostomata are far more successful, but the Agnatha contain two extant classes.
The agnathans are today represented by hagfish (Myxini) and lampreys (Petromyzontiformes), although other classes are known from the fossil record. The hagfish are the more primitive of the two groups, and it is this fact which makes it possible that modern agnathans are paraphyletic. The group comprised of all jawless fish is definitely paraphyletic, as jawed vertebrates evolved from jawless ones, but the monophyly of modern agnathans is possible, but disputed. The hypothesis stating that modern agnathans form a paraphyletic group is called the Craniata hypothesis, while the alternative is called the Cyclostomata hypothesis.
The Gnathostomata, including the majority of chordates, is composed of the microphyla Chondrichthyes and Osteichthyes (also called Teleostomi). All Gnathostomata have jaws, and are the only animals to have jaws. One group retained the ancestral cartilaginous skeleton, while the other evolved a new structure for the skeletons – bone.
The Chondrichthyes are composed of several successful groups of fish, the sharks, rays, and several other groups. The Holocephali are a small clade of cartilaginous fishes, composing within them rat fishes, rabbit fishes, and elephant fishes. The sharks (Selachimorpha) and rays (Batoidea) make up the subclass Elasmobranchii.
The following is a cladogram of the shark orders.
Source: Sharkpedia by DK Publishing, ©2008
Although the name Osteichthyes translates to "bony fish", as a clade it includes the tetrapods as well. To avoid ambiguity, another name used is Teleostomi. The bony vertebrates are made up of the ray-finned fish; lobe-finned fish; and the tetrapods, descendants of lobe-finned fish. As opposed to all other vertebrates, members of this group have bone, but retain cartilage in small portions of their skeletons.
The ray-finned fish are divided into two subclasses, the Chondrostei and the Neopterygii. The Chondrostei contains primitive ray-finned fish such as paddlefish and bichirs. This group, containing secondarily cartilaginous fish, is probably paraphyletic, with the bichirs and reedfish being basal. However, considering that the ancestral ray-finned fish most likely had a bony skeleton, the monophyly of the Chondrostei is possible. The Neopterygii is divided into two infraclasses, the Holostei and the much larger Teleostei. For more information on ray-finned fish phylogeny, see here.
The following is a simplified cladogram of the lobe-finned fish.
On land, the most dominant animals are tetrapods, a group of derived sarcopterygians. Although derived, the first tetrapods in the fossil record are found only 20 million years after the first sarcopterygians, meaning that the early lobe-finned fish were already close to being adapted to land. Most tetrapods have four limbs, each with one (in the case of equids) to five digits. The only tetrapods lacking limbs, such as caecilians and snakes have lost them secondarily. Other groups, such as cetaceans and sirens (a type of salamander) have two limbs. The four classes of tetrapods are Amphibia, Reptilia, Aves, Mammalia; through a strict cladistic view the Reptilia and Aves are combined into Sauropsida.
All amphibians go through a larval stage; the larvae, called tadpoles, lack limbs and have gills instead of lungs. Three subclasses of amphibians are known, two from fossil records and one that contains the living amphibians. Depending on the definition, as well as the cladistic view, Amphibia can be paraphyletic or monophyletic. The sensu lato view of Amphibia puts it as paraphyletic, as well as one interpretation of the Lissamphibia (Amphibia sensu stricto). Many modern phylogenetic supporters consider Amphibia vastly paraphyletic, as it usually includes all non-amniote tetrapods, and prefer a strict definition containing lissamphibia, temnospondyls, and a few others. There are however other vews that lissamphibia itself is paraphyletic, with caudates being temnospondyls, anurans in a similar spot to monophyletic lissamphibian placement, and caecalians potentially being close to the amniotes.
The labyrinthodonts are a paraphyletic fossil group of amphibians. The more derived labyrinthodonts form a clade, but included are the ancestral tetrapods, such as the oldest known tetrapod, Acanthostega and a slightly younger tetrapod (still primitive), Ichthyostega. Despite containing primitive tetrapods, it also includes the more derived Reptiliomorpha (stem-group amniotes) and Temnospondyli (possible ancestors of (certain) lissamphibians). As such, it includes both primitive and derived amphibians. Lepospondyli.
All modern amphibians belong to the Lissamphibia. It may or may not be a clade, which is discussed in the previous section. There are three order of modern amphibians, the frogs and toads (Anura), the salamanders and newts (Caudata), and the caecilians (Gymnophiona). One recent order of lissamphibians, the Allocaudata, went extinct in the Pliocene. In the original description of the Lissamphibia, the caecilians were excluded, which is accurate, as a reasonable hypothesis states that the caecilians are unrelated to the other lissamphibians. Within Lissamphibia, the Anurans and Caudates form the Batrachia, and the Allocaudates being sister to this, and Gymnophionans the most basal.
The caecilians are a legless, wormlike group of terrestrial amphibians. It is debated whether they evolved from Lepospondyli or Temnospondyli, but they are definitely the outgroup of the living amphibians. They are almost exclusively found in tropical environments, and are not well studied. Caecilians and immediate fossil relatives are placed in the clade Apoda. The caecilians are not typically divided into suborder, infraorders, and so on, but are divided into ten families. The Caeciliidae, from the Americas, is the largest family, despite having only two genera.
Allocaudata is a long lived, yet little understood order of lissamphibians that originally were thought to be caudates, are now known to be distinct. The sole known group is the Albernerpetonidae, which span from the Bathonian to the Piacenzian, but as over 95% of their probable diversity is not known, both the time range and diversity are probably/potentially far larger. Their origins are unknown, though they are possibly the sister to Anura + Caudata. They had fish-like scales in their skin, and some species are among the smallest tetrapods. Bite marks on multiple fossils indicate violence and jaw locking in fights similar to many modern salamanders. In some places, like the Wessex Formation their fossils are quite common, though often scrappy.
The salamanders and newts are a group of amphibians looking quite similar to the hypothesized ancestral lissamphibian. Unlike other modern amphibians, many species are fully aquatic. There are three superfamilies of salamanders – Cryptobranchoidea (primitive salamanders), Salamandroidea (advanced salamanders), and Sirenoidea (sirens). The Cryptobranchoidea are composed of two families, the giant salamanders (Cryptobranchidae) and the Asiatic salamanders (Hynobiidae). The Salamandroidea contains most living salamanders, and are also the most widespread group of salamanders. The Sirenoidea are unusual among amphibians in two ways. First, they have greatly reduced forelimbs and lack hind limbs. Second, the adult sirens are aquatic and retain their gills.
The Anura are composed of frogs and toads, of which there is no taxonomic differentiation. Unusual among vertebrates in general, they lack a tail. The first frogs are known from the early Triassic, although molecular studies suggest a slightly earlier origin. The frogs have more species than most other vertebrate orders. They are composed of three suborders – Archaeobatrachia, Mesobatrachia, and Neobatrachia. Out of these, the Neobatrachia is by far the largest, containing over 5,000 species. The suborders are divided into families, a list of which can be found here.
The Reptiliomorpha, or Anthracosauria, are a clade of tetrapods that include the amniotes, but exclude lissamphibians. Note that the Lepospondyli fall into this group, and therefore pottentially also the caecallians. The Diadectomorphs certainly belong here. They are the most reptile-like of non-amniotes, and many members were originally considered amniotes.
The Lepospondyli were an ancient group of diverse amphibians, and one of the more primitive groups. Many species were similar to living salamanders (a relatively unchanged group of amphibians), but others were similar to lizards and snakes. One of the most unusual of them, and likely the most famous, is Diplocaulus, with a streamlined triangular head. Their position within tetrapoda is undecided, with three different hypotheses emerging. The first states that all lissamphibians evolved from the Lepospondyli. The second hypothesis mentions Lepospondyli having no living descendants. The last considers the polyphyly of lissamphibians, with only caecilians having evolved from the Lepospondyli. (Note that this doesn't include the orders Diadectomorpha, and the rest further down the page.)
The Diadectomorphs were a group very close to the amniotes, perhaps being sister to them, for which a monophyletic version of Cotylosauria has occasionally been used as a name. They are quite morphologically diverse, with carnivorous, omnivorous, herbivorous, and perhaps piscivorous forms, some living on land, and some in water. Such diversity may indicate a fair bit of unknown diverse. First appearing in the Mississippian, and diversifying with the later climates, they disappear in the transition from Early to Middle Permian, though again this could be an illusion. In terms of general body plan, they probably wouldn't have looked out of place amongs basal sauropsids, or non-cynodont synapsids to the naked eye. The Diactids are the most well studies, represented, and the first to appear (earliest known Diadectomorph remains tentatively refered to Diadectes proper, however some ichnogenera from this group (Diadectomorpha, not Diadectidae) are signifgantly older), though whether this is an actual pattern or simply an illusion of the fossil record is unknown. There are roughly 15 taxa ("species" level) known. They are known from what is know North America and Europe.
Because of their placement, reproduction is of course debated. It has been suggested they may lie within Amniota, seperate from the synapsids and sauropsids, and therefore had amniotic eggs, but their common placement outside would lead one to think an amphibian egg would be the method of reproducing. As no eggs or tadpoles have been found, this is up to interpretation. However, the body plan and terrestriality of the diadectids and limnoscelids lead an amniotic egg to make more sense. It has been suggested that the transition to amniotics would have been more smooth then usually interpreted as, and perhaps these creatures had an intermediate, or even evolve something similar, yet different, and this is perhaps the most logical interpretation. One of the many features shared with Amniotes is their dentation, lacking the enamal and general shape of labyrinthodont-grade tetrapods. They were also the first large group of tetrapods to evolve a herbivorous lifestyle, though pareiasaurs and edaphosaurs followed in their footsteps quickly.
The following is a cladogram of the Diadectomorpha.
Amniotes are those vertebrates which are fully adapted to life on land. The ancestral amniote laid a hard-shelled egg, but several groups, especially the therian mammals, have developed different means of reproduction. Although the amniotes are masters of the land, many groups have returned to sea. These groups include cetaceans, sea cows, sea snakes, sea turtles, the marine iguana, and several extinct groups like plesiosaurs and ichthyosaurs. The amniotes are composed of two subgroups, the Sauropsida and Synapsida.
The Sauropsida is a clade of tetrapods comprising reptiles (a paraphyletic group) and birds. Sauropsids are generally synonymous with a monophyletic definition of reptiles, but under the sauropsid definition no synapsids are reptiles. Most sauropsids lay amniotic eggs, a trait not retained in mammals. However, some reptiles, especially certain snakes (like vipers), do not lay eggs, and have developed a viviparous or ovoviviparous method of reproduction. Sauropsids are divided into two groups – Anapsida and Diapsida, of which Anapsida may be polyphyletic. Depending on the place of chelonians in the Sauropsida, Anapsida may or may not have living members.
Unranked Non-Clade †Anapsida
Anapsids likely form a polyphyletic or paraphyletic group, especially considering that they traditionally include the Testudines. New research shows that chelonians are derived diapsid reptiles, so for this reason we will consider anapsids to be extinct, and will put the chelonians under diapsids. The chelonians, however, have been found to be the outgroup of living reptiles. One group of anapsids is the Captorhinida, including primitive amniotes such as Milleretta. It is likely a paraphyletic group. The Mesosauria includes very primitive aquatic reptiles, such as Mesosaurus. In some studies, they have been shown as the sister taxon to all other reptiles. The Procolophonia includes such notable anapsids as the pareiasaurs, which previously were thought to have given rise to the chelonians.
All living reptiles are diapsids, but this is not fully accepted. For this reason, I will be using a slightly modified taxonomy. Neodiapsida excludes the Testudines, which are in this taxonomy part of the Diapsida. Diapsids have two holes in their skulls, a feature which the Testudines lack. This makes is possible that the ancestral diapsid was actually an "anapsid", and that only the ancestor of the Neodiapsida had a two-holed skull. In any case, chelonians are not descendants of the traditional anapsids.
The Testudines, also called the Chelonii, are an "anapsid" group of reptiles which are actually basal diapsids. As mentioned before, this brings up the question of whether the ancestral diapsid had an anapsid skull. The chelonians comprise shelled reptiles, especially those known as turtles (sometimes referring only to sea turtles), tortoises, and terrapins. There are two suborders of chelonians, Cryptodira and Pleurodira. The Cryptodira is the larger group, comprising three superfamilies. Chelonioidea contains sea turtles, Testudinoidea tortoises and pond turtles, and Trionychia the softshell turtles. The Pleurodira, or side-nicked turtles, contain two families, Chelidae and Pelomedusidae.
Most living reptiles are members of the Neodiapsida, which includes three living orders of "reptiles" as well as birds. It can be defined as the least inclusive clade encompassing Squamata and Aves. Under this modified definition of Diapsida, the Neodiapsida are the only "true" diapsids. The Neodiapsida emcompasses the Archosauromorpha and the Lepidosauromorpha. The archosauromorphs include the archosaurs and other related diapsids, like rhynchosaurs and the famous Euparkeria. The lepidosauromorphs include Squamata, the Rhynchocephalia (of which only the tuatara remains), and the Sauropterygia (plesiosaurs and kin). The Ichthyopterygia, an important group of marine reptiles, are in the Neodiapsida, but other than that their phylogeny is unknown.
The synapsids are a group of amniotes usually referred to as "mammmal-like reptiles", even though they are not actually reptiles and the group includes mammals. The group includes the pelycosaurs (a paraphyletic group) and the therapsids. There are multiple groups of pelycosaurs, usually monophyletic, such as Edaphosauridae and Varanopidae. The most famous pelycosaur is the sail-backed Dimetrodon, a sphenacodontid. The therapsids evolved inside of the sphenacodontids, which are thus a paraphyletic group of pelycosaurs, and include many subgroups, detailed later.
Unranked Non-Clade Pelycosauria
The following is a cladogram detailing the relationships of the pelycosaurs.
Note: Therapsida likely evolved from within the Sphenacodontidae
The therapsids were a very diverse group of synapsids, united by such features as an erect posture. However, many were wiped out in the Permian-Triassic and Triassic-Jurassic extinction events, leaving the mammals as the only survivors. Of course, these groups are all very important, so it would be unfair to ignore them. The most basal group of therapsids, the Biarmosuchia, were medium-sized predators, and were all wiped out in the Permian-Triassic extinction. The next group to split off from the line leading to mammals were the Dinocephalia, including a carnivorous group (Anteosauria); and an herbivorous group (Tapinocephalia), the most famous of which is Moschops. The Anomadontia include several lesser-known groups, but most famously the dicynodonts. Other groups, closer to mammals, are the Gorgonopsia, and Therocephalia. Both of these groups are carnivorous. The cynodonts are the most mammal-like of the therapsids, and thus gave rise to mammals in the Late Triassic.
The following cladogram sums up the evolutionary relationships of the therapsids.
Mammals are a somewhat unique group of animals. They are warm-blooded, a feature only shared with birds, and have hair. They feed milk to their young, and thus all species care for their young, an actually rare reproductive strategy. It seems that some of these features are shared with their cynodont ancestors, but even cynodonts happen to be radically different from crown group mammals.
Which brings up a discussion: are mammals defined based on characteristics, or based on a crown group? It they were defined based on characteristics, all mammals would have to have fur, produce milk, and be warm-blooded. Certain derived cynodonts, like Morganucodon, would probably be considered mammals under the characteristic definition, but are non-mammalian cynodonts based on crown group. Even if Morganucodon was a direct ancestor of mammals, it would still have lived before the split of Prototheria and Theriiformes, and not be part of Mammalia under the definition of the least inclusive clade containing Homo sapiens + Ornithorhynchus anatinus. Still, no consensus has been reached on the definition of Mammalia.