This is a _big_ list of vestigial features in various
organisms; I include not only animal-kingdom examples, but also plants
and even cells. Much of the material listed below can be found in
various standard reference works .Anything that discusses anatomy,
especially comparative anatomy, is good. Embryology is also a good
place to look. Other places: Charles Darwin's _Origin of Species_
(even if a bit dated :-), Stephen Jay Gould's books, and Douglas
Futuyma's _Science on Trial: the Case for Evolution_.


*** Vertebrates

* The wings of flightless birds. These have all the anatomical
features of the wings of flying birds, at least to the extent that
they are developed at all. In some species, like ostriches, they are
display organs, while in water birds, such as penguins, they are used
as fins while underwater, while in some cases, as in kiwis, they have
no function whatsoever, representing a reduction in the number of
usable limbs from 4 to 2.

* Expressible bird-tooth genes. No present-day bird or Cenozoic fossil
bird is known to have teeth ("hen's teeth" are a proverbially
nonexistent item), though Mesozoic birds including _Archaeopteryx_ did
have teeth. However, some work in the 1980's by someone named Kollar
suggests that the genes are still there. He put some embryonic chicken
jaw tissue next to some embryonic mouse/rat's jaw tissue in a mouse's
eye, and the chicken jaw tissue grew teeth that were peg-shaped with
conical tops like Mesozoic bird teeth -- and unlike most rodent
teeth. However, some recent work has suggested that this experiment
has been a false alarm, in which case it does not belong on this list.

* Most mammal tails. These are usually very reduced in comparison with
the rest of their owners' bodies, unlike the case of many fish,
amphibians, and reptiles, past and present, whose tails are continuous
or nearly so with their trunks. However, some of them do have limited
functionality. Horses' tails are fly-whisks, the bushy tails of
squirrels, anteaters, and some others may serve as false bulk to
deceive predators (picture a hawk swooping down on a squirrel's tail
instead of the rest of the squirrel), and sirenians and cetaceans have
re-evolved thick tails for propulsion. But in a lot of cases, tails
are only display organs (dogs and cats, for example), they have no
apparent function, or they are lost altogether.

* Connection to having a brachiating ancestor? In the case of the
apes, loss of the tail probably came about as an adaptation to
brachiating (gripping branches and hanging beneath); the tail may have
been too much of a nuisance (it may also have been that for frogs,
which are hoppers). An indicator of a brachiating heritage may be our
ability to stick our arms straight upward, even though brachiating is
not a common form of locomotion in our species.

* Stumpy tails and other such features of some domestic animals bred
to have none.

* Embryonic tails of tailless animals. This is universal; even human
embryos have tails at one point. In our species, all that is left is
the coccyx, a tiny bone attached to the end of the spinal-column part
of the pelvis. In the case of frogs and toads, which have tiny tails
(if any at all), tadpoles have big and functional tails -- and live
like fish in the water.

* Tadpoles. Immature frogs go through this phase, in which they look
and act much like fish. Also, immature land salamanders are typically
aquatic.

* Gill bars in the embryos of tetrapods (land vertebrates and their
aquatic descendants). Though they never become functional gills, they
do get used for other things. These structures are an indicator of
piscine heritage.

* Lengthening limb bones, extra toes and the like.

	The standard tetrapod limb configuration goes like this: two
or three bones in the trunk, a single limb bone, then two limb bones,
then a mass of small bones, then several (often 5, but sometimes
fewer) strings of bones in the digits. In early embryos, the limbs are
short and stubby, and all of the bones are short and cubey; they have
a clear resemblance to fish-fin bones in that respect. However, in
many land animals, the first single (humerus, femur) and second pair
of bones (radius+ulna, tibia+fibula) often grow long; the digit bones
sometimes grow long also, while the hand/foot bones in between stay
small -- but generally unfused.

	There are numerous cases of superfluous limb bones. Dogs and
cats and their relatives have separate sets of digit bones, but their
paws are solid units, with only a bit of a split at their
ends. Furthermore, dogs (at least) have "dewclaws" on digits a bit up
the limbs -- which are useless. Most even-toed ungulates (cows,
antelope, sheep, goats, deer, pigs, etc.) have on each foot two
load-bearing digits with hooves -- and two small digits behind
them. Furthermore, their hooves sometimes look like a single hoof
split in the middle, as if there is no special reason for there being
two. Horses and their living relatives (donkeys, zebras) have a single
load-bearing, hoofed, digit, but they often have two toe bones
("splints") on each side. In mid-Cenozoic equines, however, these
extra toe bones were much more prominent, and present-day equines are
sometimes born with extra toes on each side of their load-bearing
toes. Elsewhere, birds have an Alula, or "bastard wing" on their
wings, which is an extra digit.

* Human toes. Our feet have very small toes which are not good for
very much. Furthermore, one toe is much larger than the others and has
a noticeable gap between it and its neighboring toe. In our nearest
simian relatives, chimpanzees and gorillas, the toes are longer, and
the big toe often points outward, like a thumb. In some early hominid
species, the toe bones are, in fact, longer than in our species.
	
* Fused bones. There are numerous examples of bones that start out
separate, but get fused later on. There are numerous examples. The
pelvis is a single bone formed from the fusion of the hindlimb
equivalent of the forelimb's shoulderblade and collarbone -- and some
vertebrae, which still keep a lot of their shape. The skull starts off
as several bones, which usually fuse as its owner grows. The wishbone
of birds is the two collarbones fused together. Likewise, there is a
lot of bone fusion in the outer wing bones (wrist + digits). Frogs
have the paired limb bones (radius+ulna, tibia+fibula) fused
together. Some dinosaurs had a long string of fused lower vertebrae,
making a rigid tail.

	Are some teeth fused? Mammalian back teeth (premolars,
molars), have complicated upper surfaces and multiple roots,
suggesting that they are derived from the fusion of simpler, peglike
teeth (which reptiles have).

* Giraffe neck lengths. Baby giraffes start out with necks whose
relative length is similar to those of other ungulates; it is as they
grow that they acquire the relatively long necks that the species is
noted for. These necks have the usual mammalian number of 7 vertebrae
-- 7 long vertebrae.

* Solid-color equines (horses and donkeys) having offspring with zebra
stripes. This seeming perplexity can be explained if the original
equines had genes for making stripes that were preserved in zebras,
but got switched off in other equines. Reshuffling of chromosomes in a
hybrid may lead to the stripe-making genes getting switched on again.

* Fetal teeth missing from adults. Fetuses of cows and other ruminants
have upper front teeth forming in their jaws; these are later
resorbed. Baleen-whale fetuses also have un-erupted teeth which are
later resorbed; these are cone-shaped, like the teeth of toothed
cetaceans. Anteaters also go through this developmental detour.

* Snakes with rudimentary limbs. Boa constrictors have small hind
legs; these may aid in copulating. However, most other species of
snakes lack this feature.

* Snakes with one lung small. In many snake species, only one lung
stays big; the other one becomes very tiny, but does not
disappear. Some snakes, however, do have equal-sized lungs (the usual
tetrapod situation).

* Cetacean hipbones. Some whales have hipbones deep inside their
bodies, attached to no limbs. One possible purpose is to serve as an
attachment point for muscles that move a male cetacean's penis,
however.

* Moving eyes. In early vertebrate embryos, the eyes are on the sides
of the head, and they stay that way in most species. However, in some
groups (primates, carnivores, and owls), the eyes move from the sides
to the front of the head, making possible binocular vision. Since
these groups are all surrounded by side-eye animals, it is clear that
binocular vision evolved at least three times.

	Flounder eyes. On sea floors, there live certain fish that lie
on their sides. They have two eyes -- on one side of their heads. But
they start off life with eyes on both sides of their heads, and one
eye moves to the other side. In this and the previous case of moving
eyes, why is it necessary for that to happen?

* Wisdom teeth. Our jaws are a bit small for these late-erupting
teeth; some people have them, while others do not. And some people
(like myself) have two upper wisdom teeth but no lower wisdom teeth.

* Outsized hind legs of some four-legged dinosaurs. _Stegosaurus_, for
example, had hind legs much bigger than its front legs, desite it
walking on all fours. This is probably a byproduct of being descended
from a two-legged ancestor that went back to walking on all
fours. Many of the dinosaurs walked on their hind limbs only, with the
front limbs remining at various levels of development. In
_Tyrannosaurus_, they are _very_ small, though still there, which has
led to the suggestion that they are vestigial. The earliest dinosaurs
known, like _Herrerasaurus_, had walked on two cases. Transitional
cases? Possibly! _Iguanodon_ or some other such dinosaur apparently
walked on two legs when juvenile, and on all fours when adult (and a
lot heavier).

* The Hoatzin chick's claws. The claws on their wing limbs enable them
to climb away from potential predators; their presence indicates that
all the clawless-winged birds have the potential of growing claws on
their wing limbs, which is inherited from their clawed-limbed land
ancestors, which were probably small theropod (carnivorous) dinosaurs.

* Hollowness of dodo and penguin bones. It is not critical for ground
birds to reduce weight with hollow bones of the sort that flying birds
have.

* Gill bars of tetrapod embryos. The cartilage gill bars appear, only
to disappear or be reworked into other features with later growth. Of
these animals, only amphibians have gills, and that only in the larval
(tadpole) stage. Most adult amphibians and all the rest are air
breathers; even the aquatic ones do not grow gills to use underwater

* Jaw origins from gill bars. In jawed-vertebrate embryos, the jaws
are formed from the gill bars closest to the mouth. In jawless fish
(lampreys and hagfish), these gill bars stay gill bars. This
circumstance indicates an origin of jaws from gill bars.

* The mammalian amniotic sac. This is a vestigial eggshell that
surrounds the fetus. Live birth clearly evolved out of retaining an
egg inside (which some animals do).

* Aquatic-tetrapod air breathing and breeding on land. Aquatic animals
like sea turtles, Galapagos iguanas, sea snakes, crocodilians, water
birds (penguins, for example), pinnipeds (seals, sea lions, and
walruses), sirenians (manatees and dugongs), and cetaceans (dolphins
and whales) all have to come up to the surface to breathe, which is a
serious limitation for an aquatic animal. And of these, only the sea
snakes, sirenians, and cetaceans are completely aquatic, giving birth
in the water; the rest either lay eggs or give birth on land. Needing
to breathe air is perhaps most striking for cetaceans, which are
otherwise very well-adapted for life in water, with their being
fish-shaped and excellent swimmers.

* Lesbian lizards. Parthenogenetic lizards (_Cnemidophorus_) are all
one sex, but they need to copulate in order to lay eggs. An older one
acts like a male and a potential egg-layer acts like a female, and the
"male" bites the "female" in the neck to make it lay eggs.

* Moving testes. [in response to a posting of some examples of bad
design] ...the even worse design of having the testes form inside the
abdomen, then have to pass through the abdominal wall and down to the
scrotum, thereby leaving a weak spot (two, actually) in the wall.
This spot, called the inguinal canal, can herniate, allowing the
intestines to slop out under the skin.  Herniation both screws up the
intestine and cuts off/slows the blood flow to the affected
testis. Great design [contributed by Paul Keck]. This is a general
mammalian feature; however, most other vertebrates have their testes
firmly inside the abdominal wall.


*** Invertebrates:

* Homeotic [One Part Like Another] Mutations in Insects.

	Fruit flies sometimes grow legs instead of antennae on their
heads, and mosquitoes sometimes have legs instead of mouthparts. At
first sight, it might not seem as if a single mutation can be
responsible for such multiple-feature effects. However, if legs,
mouthparts, and antennae originated as modifications of a single kind
of ancestral appendage, then this kind of mutation would become more
understandable. It could be that the default growth pattern for an
appendage is to become a leg; a default that would be overridden by
control proteins made from genes expressed only in front segments. If
the genes for those control proteins had mutations that made the
proteins defective, then these appendages would develop in default
fashion -- as legs and not as antennae or mouthparts (it's not even
that simple because legs on different segments often develop
differently).

	There is another such mutation that causes a fly to start to
grow legs on its abdominal segments; but that mutation kills the fly
while it is still a maggot. It is almost as if all the abdominal
segments could potentially grow legs, but are not. Looking among the
other arthropods (centipedes and millipedes, arachnids, crustaceans,
etc.), there are several different configurations of limb
specialization and disappearance, including the case of all segments
having lookalike limbs. That particular one is found in trilobites,
which go back to the base of the Cambrian, which is when hard-shelled
marine animals first emerged. Several other arthropods of that time,
as in the Burgess Shale animals, had had that sort of limb
configuration. In fact, it is very likely that the ancestral arthropod
was like that, with many of its descendants having their limbs
specialized in different fashions to become antennae or mouthparts or
whatever -- or even suppressed. So it is logical to conclude that the
aforementioned mutations represent throwbacks to the limb
configurations of less specialized ancestors.

* Homeotic Wing Mutations: These kinds of mutations also occur in wing
structure. Most (living) insects have two pairs of wings, one per
segment, but dipterans (flies and mosquitoes) have only one pair, the
front one. Instead of the rear one, there is a pair of "halteres"
(balancing organs). However, there are fly mutations known that
produce wings instead of halteres,. There are also cockroach mutations
that produce wings on the next segment behind (a third pair), and
there are some fossil insects with three pairs of wings (one for each
thorax segment).

* Small wings of the flightless females in certain moth species. In
most insect species, both sexes are fliers.

* Crab tails. These are small and tucked under the animal's body; they
are very likely a hangover from lobsterlike ancestors, who had had a
more typical arthropod body shape.

* Torting (twisting) of gastropods (snails, etc.) as they go. Early in
their growth, their body gets a 180-degree twist, putting the anus to
the right and then to the top, nearly above the head. That is good for
living inside a shell, where one would not want to excrete, but some
gastropods that have lost their shells (slugs) tort -- then de-tort --
as they grow. A result of this torting is that right-hand-side organs
are reduced; this sometimes remains after slugs' de-torting. [Example
contributed by Chris Colby (colby@bu-bio.bu.edu) and Matthew P Wiener
(weemba@sagi.wistar.upenn.edu)].


*** Plants:

* Flowers of Self-Pollinators: Dandelions and some other asexually
reproducing plants still make flowers, although they have no need to
attract pollinators.

* Modified leaves: The petals and sepals of flowers are simply
modified leaves. Stamens and pistils are also modified leaves, though
much more extensively modified ones.

* Vestigial flower parts: Some non-flowering angiosperms, like the
grasses (which are wind-pollinated), have small structures that are
almost certainly vestigial flower parts.

* Nonfunctional pistils in male flowers: Some plants have separate
flowers of each sex (sometimes on the same plant, and sometimes on
different plants). But since the predominant configuration of flowers
is to have both sexes of reproductive organs (stamens and pistils),
the pistil of a flower with only stamens functional is vestigial.

* Alternation of Generations. Many algae and "lower" plants, like
mosses and ferns, have an alternation of generations between an
asexual diploid phase and a sexual haploid phase. In ferns and similar
plants, it is the diploid phase which is the most prominent; it
reproduces by producing spores. The haploid plants are small ones that
release egg and sperm cells; they need damp ground for the sperms to
swim to the eggs in, thus limiting ferns' habitats. Looking at the
"higher" plants, the gymnosperms and the angiosperms, we find that
just about all of the plant is the diploid phase. The female haploid
phases grow in the reproductive organs of the diploid phases; they are
only a few cells in angiosperms. The male haploid phases are released
as pollen; when they alight on a pistil, they sprout a tube that grows
into it in order to get to the female haploid phase inside.

* Leaves of Parasitic Plants. Some plants are saprophytic (living on
dead material, fungus-style) or parasitic; but some of them still grow
stems and leaves, despite their not using them for photosynthesis
(they are white and not green). They may also have nonfunctional,
pigmentless chloroplasts.


*** Cells:

* Gene Duplications and Pseudogenes:

	There is an abundance of evidence for gene duplications. This
evidence comes out of finding family trees for genes and proteins;
sometimes, several sequences from the same organism will come out
similar enough to suggest common ancestry. For example, hemoglobin in
the configuration found in most vertebrates (though not the lamprey)
has four amino-acid chains, two "alpha" and two "beta". The two have
parallel family trees, and the family tree of both of them has a
divergence between the alphas and the betas a little below their first
divergences -- between sharks and the group bony fish + land
vertebrates. So there may have been only one hemoglobin gene (as in
what lampreys still have) that got duplicated in some early fish; the
duplications being carefully preserved in all that fish's descendants.

	There is also abundant evidence of "pseudogenes", gene
sequences that are not expressed because they do not have a "start"
signal. They may have originated the same way that duplicated genes
did, but a mutation that disabled the "start" signal would not be
selected against because there would be a functional copy of a gene
that did what it does.

* Mitochondria and Plastids (Chloroplasts, etc.) in Eukaryotic Cells
(Those with Well-Defined Nuclei) -- the Endosymbiosis Hypothesis:

	The large majority of eukaryotic cells contain mitochondria,
and many of them contain plastids. Mitochondria perform respiration
(combining food molecules with oxygen), while plastids perform
photosynthesis (making food molecules from inorganic compounds). These
structures have several features which suggest that they are descended
from free-living ancestors -- and there are even good candidates for
their closest free-living relatives. These features are:

	Separate genomes and protein-synthesis systems (ribosomes,
transfer RNA, etc.). This redundancy is especially curious since (1)
many of their proteins, especially for mitochondria, are coded for in
the nucleus and (2) mitochondria, at least, have a significantly
higher mutation rate than nuclei, which suggests that keeping genes
there is risky. Copies of the transferred genes may have drifted
across membranes to nuclei, and preserved there because the copying
there is more reliable. Those remaining may remain because they code
for proteins that are awkward to transport inward.

	Multiple membranes. Mitochondria are surrounded by not one,
but two, membranes, and plastids are surrounded by two, three, or four
membranes. This is a curious oversupply, since only one is necessary
to separate two regions, and since more than one represents that many
extra barriers for molecules to cross (most of the biochemical action
happens at or inside the inner membrane). However, it is to be
expected of an once-free-living organism living in a bubble created by
the host cell. For some plastids, it appears that endosymbiosis
happened more than once; for four-membrane plastids, there is a
"nucleoid" between the second and third membranes, which is probably a
vestigial eukaryotic nucleus.

	Membrane topology. Cell membranes often have a well-defined
"inside" and "outside", the "inside" having more protein complexes and
the like sticking out. For 2-membrane organelles, the topology going
outward is: in/out, out/in; for 3-membrane organelles, it is in/out,
out/in, out/in, and for 4-membrane organelles, it is in/out, out/in,
in/out, out/in. This is all consistent with the endosymbiosis
hypothesis; the last two cases are the signature of two such events.

	Close relations. This can be found by comparing biochemical
details, like gene sequences for critical molecules such as ribosomal
RNA and cytochrome c. The closest relatives of the mitochondria turn
out to be the Purple Photosynthetic Bacteria, which have one-step
photosynthesis working from sulfur compounds or organic molecules, and
their nonphotosynthetic relatives, many of which are oxygen-users (a
candidate for the closest: _Paracoccus_). Among them are numerous
familar Gram-Negative bacteria, including _Escherichia coli_ and
root-nodule bacteria. Plastids, on the other hand, are most closely
related to the Cyanobacteria (Blue-Green Algae); both have two-step
photosynthesis that releases oxygen from water (the "steps" are
photon-absorbing steps). The nucleus, however, has a more obscure
genealogy; the best candidates so far for its closest relatives are
some obscure wall-less (with only a flexible membrane) bacteria such
as _Sulfolobus_ and _Thermoplasma_, that like to live in hot
springs. These are only distantly related to the bacteria discussed
above ("eubacteria", which include the examples mentioned above and
most other familar bacteria), and fall into a group called the
"archaebacteria", which include methanogens (energy source: hydrogen +
carbon dioxide -> methane + water) and extreme halophiles (which
sometimes color salt ponds).

	How plausible? There are numerous present-day examples of
symbioses analogous to the hypothesized endosymbioses; one of them is
lichen, which is an alga with a fungus, each of which does what the
other cannot.

* More such wayward microbes?

	Hydrogenosomes are structures in some anaerobic protists
(eukaryotic one-celled organisms) that release hydrogen as part of
their energy production for the cell. These have two membranes, but no
genome. It is suggested that they are mitochondria that have lost the
ability to use oxygen, and just release hydrogen. They would have lost
the genes for the necessary molecules, and probably the only
"rationale" for keeping a genome.

	Eukaryotic flagella/cilia. It has been suggested that these
and other such structures, such as the centrioles that are involved in
a dividing nucleus, are originally symbiotic bacteria such as
spirochetes (some protists have lots of spirochetes attached to them);
but that is a far-from-settled question.

* Oxygen Use: A Later Addition?

	There is a remarkable feature of the use of oxygen; it is that
most of the reactions that use it either release it as a first step or
use it as a last step. Just about all of the reactions in between (and
some biosynthesis reactions can use a large number of steps) proceed
as if oxygen never existed. There are some exceptions, like a step in
steroid synthesis, but these are rare. This would be natural if oxygen
metabolism was a latecomer, and built onto existing oxygen-less
metabolic processes.

	This conclusion is confirmed by molecular-evolution studies;
aerobic lineages (those that use oxygen) are often surrounded by
anaerobic ones, and the exact molecules used in respiration (oxygen
use for energy) show a lot of variation. These facts are consistent
with the hypothesis of multiple acquisition of oxygen use after oxygen
became abundant about 2 billion years ago. Around that time, rocks
start becoming more oxidized; U3O8 replace UO2, and there are big
deposits of Fe2O3 (rust in Banded Iron Formations); the previous FeO
was more soluble.

	Furthermore, bacterial photosynthesis has two main forms. The
simpler, and more widespread form, uses only one photon-absorption
step, with only one photosystem. Cyanobacteria and chloroplasts use
two-step (two-photosystem) photosynthesis and release oxygen; the
second photosystem may have originated by duplication of the genes for
the first one (gene duplication is a common event in molecular
evolution). So O2 release, as well as O2 use, was almost certainly a
later add-on.

* The RNA World (speculative)

	The question of how a complex, interdependent system of
interacting nucleic acids and proteins could have evolved from scratch
now has a possible answer: the "RNA World" hypothesis, which states
that the original organisms had genes and enzymes composed of
RNA. This was inspired by the discovery of "ribozymes" -- RNA strands
that can act as enzymes. In this hypothesis, DNA is a derivative of
RNA that was invented to store genetic information in a medium
distinct from RNA, so it would not be eaten by RNA-recycling enzymes,
and proteins are also a later invention, for the purpose of making
more efficient enzymes.

	Here are some features at least cosnsitent with the RNA World:
Ribosomes (the assembly sites for proteins) have both RNA and
proteins; they can function without the proteins but not without the
RNA. There are some coenzymes (small molecules that work with
enzymes), such as NAD (niacin) that look like small bits of modified
RNA, and the energy-storage molecule ATP contains an RNA nucleotide
(adenosine). The energy of ATP resides in its pyrophosphate
(phosphate-phosphate) bonds; the adenosine is almost certainly a
"handle" for the enzymes that work with it. Could the RNA nucleotides
of many coenzymes also be "handles"?


Some Refs to This Section:

	_Bacterial Evolution_, C.R. Woese, Microbiological Reviews,
Vol. 51, No. 2, p. 221; June 1987

	_Archaebacteria_, C.R. Woese, Scientific American, 1987(?)

	_The Phylogeny of Prokaryotes_, G.E.Fox et al. (including C.R.
Woese), _Science_, Vol. 209, p. 4455; July 25, 1980