Content uploaded by Matan Shelomi
Author content
All content in this area was uploaded by Matan Shelomi on Oct 25, 2015
Content may be subject to copyright.
Annals of Improbable Research | July–August 2012 | vol. 18, no. 4 | 15 www.improbable.com
A P
E H
P
by Matan Shelomi1, Andrew Richards1, Ivana Li1,
Yukinari Okido2
With the phylogenetic and evolutionary relationships of the kingdoms
Animalia, Plantae, and Fungi mostly out of the way, attention is
now turning towards the Monstrasinu, commonly known as “Pocket
Monsters” or “Pokémon” for short. Starting from the 151 original
“species” described by Japanese scientist Satoshi Tajiri in a 1996
monograph,Pokémonsciencetodaycontinuestobearewardingeld
for taxonomists. Every three to four years, several new species are
discovered and described almost simultaneously. A total of 646 Pokémon
have been described, most of
them in Japan.
Thispaperrepresentstherstattempttocreateaquantitativephylogeny
of the Pokémon, using the underlying assumption that Pokémon evolved
via natural selection independently from the animals and plants
more familiar to Western zoologists. The goal was to apply modern
evolutionarytheoryandtechniquestoaeldpreviouslylimitedtopre-
Darwinian methods of inquiry.
e Need for a Taxonomy
Conservationists have highlighted the importance of documenting extant
Pokémon, many of which are known only from single specimens and all
ofwhicharethreatenedbythePokémonghtingringsthataregrowing
rapidly in popularity, particularly among urban youth.
Further exploration into the world’s Pokémon diversity has been
limited in part due to disagreements over which species concept to
use for the Pokémon, as translations of Tajiri’s orginal work used the
term “evolution” to describe what should otherwise have been called
“metamorphosis.” Several sexually dimorphic taxa have had males and
femalesidentiedasseparatespecies(ex:NidoqueenandNidoking),
yet these errors remain as Pokémon taxonomists have shown complete
resistance to revising the Pokémon taxonomy.
Further complicating the issue is the fact that Pokémon are quite willing
to interbreed successfully: the lack of post-zygotic reproductive isolation
is one thing, but how a 400--kilogram Wailord is able to mate with an
11-kilogram Skitty at all remains a mystery. The results of the mating
are, in at least one respect, puzzling. To our knowledge, no hybrids are
created; the interbred offspring are always the same species as the female
parent, yet with some traits inherited from the male.
1 Department of Entomology, University of California, Davis, USA
2 Oak Pokémon Research Laboratory, Masara Town, Japan
Figure 1. The phylogeny of all the known Pokémon families. Sexually
dimorphic species are labeled with a *. Taxon pairs that represent two
possible metamorphosis endpoints for a single species (Ex: Slowpoke
can become a Slowbro or a Slowking) are labeled with a **, and if
they also represent sexual dimorphisms are labeled with ***.
continued >
16 | Annals of Improbable Research | July–August 2012 | vol. 18, no. 4 www.improbable.com
Methods
Wild Pokémon were captured by undergraduate, high school,
and primary-school aged interns/trainers from the Kanto,
Johto, Hoenn, and Sinnoh regions of Japan, as well as Unova
county, New York State. [Disclosure: Trainers may or may
not have used their Pokémon for combat during the course of
their research.]
For the purposes of the phylogeny, each “taxon” in the study
was a single Pokémon “family” (e.g. Pichu, Pikachu, and
Raichu are a single taxon). Sexually dimorphic species were
kept as two separate taxa as per the traditional taxonomy.
Pokémon genetics and molecular biology remains poorly
understood, so our cladistic analysis used the following
shared characters: Type (e.g. Fire, Water, Bug, Ghost,
Fighting, Steel, etc.), Egg group (16 categories limiting
which hybridizations are possible), Body Style (14 categories
describing general morphology), and moves and abilities.
Over 700 moves and abilities are known, and whether a
given taxon was capable of learning a move via natural
development was the main synapomorphy (shared, derived
characteristic) used in our phylogeny. Detailed descriptions
ofthesecharactersandtheirofcialvaluesforeachtaxonare
availableatonlinePokémonencyclopediassuchasthefth
generation National Pokédex1andpractical,ghting-and
breeding-oriented databases2.
Thedatawassavedasaspace-delineatedleandconverted
toaNexusleusingtheprogramMesquite.Thesoftware
MrBayes3 was used to run a Bayesian MCMC (Markov chain
MonteCarlo)analysisofphylogeny.Thenaltree(g.1)
was a result of 16 million generations of simulated Pokémon
evolution.
Results
The tree, rooted at its midpoint, suggests the following
evolutionary history: Pokémon life began in the water,
withPokémonsimilartolampreysandbonyshesbeing
among the earliest to reach their present state. Terrestrial life
arose independently three times, once with the evolution of
the monophyletic Ice types (starting with the semiaquatic
Dewgong), and once with the evolution of a major clade of
Flying types (starting with the seabird Pelipper). Psychic
types are a monophyletic group from within the birds,
startingwithXatu,thatgraduallylosetheabilitytoy
(suggestingLevitationismodiedight).Thehighly
humanoid Mr. Mime appears here, as the most derived and
recently evolved taxon. Also appearing here are Pokémon
suchasMewtwoandArceus,ofwhichthereareunveried
rumors of high cognitive ability, on par or greater than
humans. Certain religious sects revere the Pokémon in this
clade as deities.
Terrestriality evolves a third time, again among the water-
types, and the rest of the Pokémon evolve from this ancestor.
Theserst,fullyterrestrialPokémonarepredominantly
Normal types, and specialized types such as Grass, Fire,
and Electric arise as monophyletic groups at different
points, many from clades analogous to similar organisms in
Animalia. For instance, the Fire types arise from Pokémon
with characteristics similar to canines, and the Grass types
from hoofed Pokémon similar to ungulates. Steel and Dragon
types are both monophyletic groups within the Rock and
Ground clade. Fighting and Bug types are sister groups,
whose common ancestor is shared with the Ghost/Dark
Clade; and their common ancestor is shared with many of
the Poison types. Overall, Pokémon of the same type were
grouped together, with their moves and abilities determining
their position within each group. Egg groups and Body
Styles are completely jumbled throughout the tree.
Thebranchlengthsonthetreeindicatetime.Nospecic
units could be determined, but the relative lengths can be
compared. Many of the highly specialized Pokémon with
1 http://www.pokemon.com/us/pokedex/
2 http://bulbapedia.bulbagarden.net/
3 http://mrbayes.sourceforge.net/
Annals of Improbable Research | July–August 2012 | vol. 18, no. 4 | 17 www.improbable.com
unique move-sets show a long period of time since their
most recent shared ancestor. We hypothesize that transition
species between such pairs may await discovery. Long time
spans are also seen in cases of convergent evolutions of
elemental types independent of their main groups (labeled in
g.1withblackbarswithincoloredclades).Anexampleis
the evolution of Electrode, a pure Electric-type, among the
Steel types with its sister taxon, Magnezone, a dual-typed
Steel and Electric type. This presents convergent evolution
of Electric-type moves and abilities with loss of Steel-type
attributes. While it seems improbable that the ability to
generate and manipulate electricity across open air could
have arisen more than once, evidence suggests it indeed
happened in this case, along with many similar situations
across the tree.
Discussion
The tree seems to support most hypothesized relationships
between the Pokémon taxa, such as the close relationship
betweenMewandthegeneticallymodiedPokémon,
Mewtwo, or the monophyly of several of the Legendary
Pokémon groups (ex: Registeel, Regice, and Regirock; Ho-
oh and Lugia; or Reshiram and Zekrom). Sexually dimorphic
“species” almost always appeared as sister groups, further
supporting their lumping as single taxa. Morphologically
similar Pokémon also often appear as monophyletic (e.g.
the three feline Pokémon, Persian, Purugly, and Liepard).
The unusual Pokémon Ditto, which is its own egg group but
can successfully breed with any other Pokémon, retains its
uniqueness. The Grass-type Pokémon, which spurred furious
debate over whether they should be considered Plants or
Animals, are shown to be a monophyletic group that evolved
from a clade of Normal-type quadrupeds; the half-plant, half-
reptile Venusaur appears as the transitional species between
these two stages.
The biological species concept does not seem to apply to
the Pokémon. Monophyletic groups of Pokémon are more
consistently similar to the Pokémon Types than the breeding-
related Egg Groups. This jarring disconnect suggests that
the transmission of character traits through generations
in Pokémon does not happen through the Mendelian
genetics we are familiar with today. This paper thus sheds
considerable doubt on whether Pokémon use DNA to
transmit genetic information, and further suggests the
Monstrasinu are a unique domain of life.
Acknowledgements:
Pokémon and all that goes with it are owned by Nintendo
Co., Ltd. Dr. Okido would like to thank his collaborators,
Dr. Utsugi of Wakaba Town and Dr. Nanakamado of Masago
Town. He also wishes to thank his dedicated, unnamed
graduate student.