The ability to reliably perceive the emotions of other people is vital for normal social functioning, and the human face is perhaps the strongest non-verbal cue that can be utilized when judging the emotional state of others (Ekman, 1965). The advantages of possessing this ability to recognise emotions, i.e., having emotional intelligence, include being able to respond to other people in an informed and appropriate manor, assisting in the accurate prediction of another individual’s future actions and additionally to facilitate efficient interpersonal behavior (Ekman, 1982; Izard, 1972; McArthur & Baron, 1983). In the current experiment the consistency with which emotions display by a human female face and a Pokémon character are investigated.
The current study employed 30 hand drawings of Pikachu, a first generation electric-type Pokémon character, depicting a range of emotions (images used with permission from the illustrator, bluekomadori [https://www.deviantart.com/bluekomadori]; based on the video game characters belonging to The Pokémon Company); see Fig. 1a for examples. Also, 30 photo-quality stimuli displaying a range of emotions, expressed by the same female model, were taken from the McGill Face Database (Schmidtmann et al., 2016); see Fig. 1b for examples. Ratings of arousal (i.e., the excitement level, ranging from high to low) and valence (i.e., pleasantness or unpleasantness) were obtained for each image using a similar method to Jennings et al. (2017). This method involved the participants viewing each image in turn in a random order (60 in total: 30 Pikachu and 30 of the human female from the McGill database). After each image was viewed (presentation time 500 ms) the participants’ task was to classify the emotion being displayed (i.e., not their internal emotional response elicited by the stimuli, but the emotion they perceived the figure to be displaying).
The classification was achieved via “pointing-and-clicking” the corresponding location, with a computer mouse, within the subsequently displayed 2-dimensional Arousal-Valence emotion space (Russell, 1980). The emotion space is depicted in Fig. 1c; note that the red words are for illustration only and were not visible during testing, they are supplied here for the reader to obtain the gist of the types of emotion different areas of the space represent. Data for 20 observers (14 females) was collected, aged 23±5 years (Mean±SD), using a MacBook Pro (Apple Inc.). The stimuli presentation and participant responses were obtained via the use of the PsychToolbox software (Brainard, 1997).
Figure 1. Panels (a) and (b) illustrate 3 exemplars of the Pokémon and human stimuli, respectively. Panel (b) shows the response grid displayed on each trial for classifications to be made within (note: the red wording was not visible during testing). Panels (d) and (e) show locations of perceived emotion in the human and Pokémon stimuli, respectively. Error bars present one standard error.
The calculated standard errors (SEs) serve as a measure of the classification agreement between observers for a given stimuli and were determined in both the arousal (vertical) and valence (horizontal) directions for both the Pokémon and human stimuli. These are presented as the error bars in Fig. 1d and 1e. The SEs were compared between the two stimulus types using independent t-tests for both the arousal and valence directions; no significant differences were revealed (Arousal: t(58)=-0.97, p=.34; and Valence: t(58)= 1.46, p=.15).
Effect sizes, i.e., Cohen’s d, were also determined; Arousal: d=0.06, and Valence: d=0.32, i.e., effect sizes were within the very small to small, and small to medium ranges, respectively (Cohen, 1988; Sawilowsky, 2009), again indicating a high degree of similarity in precision between the two stimuli classes. It is important to note that the analysis relied on comparing the variation (SEs) for each classified image (reflecting the agreement between participants) and not the absolute (x, y) coordinates within the space.
What could observers be utilizing in the images that produce such a high degree of agreement on each emotion expressed by each stimulus class? Is all the emotional information contained within the eyes? Levy et al. (2012) demonstrated that when observers make an eye movement to either a human with eyes located, as expected, within the face or non-human (i.e., a ‘monster’) that has eyes located somewhere other than the face (for example, the mythical Japanese Tenome that has its eyes located on the palms of his hands; Sekien, 1776) the observers’ eye movements are nevertheless made in both cases towards the eyes, i.e., there is something special about the eyes that capture attention wherever they are positioned. Schmidtmann et al. (2016) additionally showed that accuracy for identifying an emotion was equal when either an entire face or a restricted stimulus showing just the eyes was employed. The eyes of the Pikachu stimuli are simply black circles with a white “pupil”, however they can convey emotional information, for example, based on the positions of the pupil, the orientation of the eye lid, and by how much the eye is closed. It is hence plausible that arousal-valence ratings are made on the information extracted from only the eyes.
However, for the Pokémon stimuli Pikachu’s entire body is displayed on each trail, and it has been previous shown when emotional information from the face and body are simultaneously available, they can interact. This has the result of intensifying the emotion expressed by the face (de Gelder et al., 2015), as perceived facial emotions are biased towards the emotion expressed by the body (Meeren et al., 2005). It is therefore likely that holistic processing of the facial expression coupled with signals from Pikachu’s body language, i.e., posture, provide an additional input into the observers’ final arousal-valence rating.
Whatever the internal processes responsible for perceiving emotional content, the data points to a mechanism that allows the emotional states of human faces to be classified with a high precision across observers, consistent with previous emotion classification studies (e.g., Jennings et al., 2017). The data also reveals the possibility of a mechanism present in normal observers that can extract emotional information from the faces and/or bodies depicted in simple sketches, containing minimal fine detail, shading and colour variation, and use this information to facilitate the consistent classification of the emotional states expressed by characters from fantasy universes.
Brainard, D.H. (1997) The psychophysics toolbox. Spatial Vision 10: 433–436.
de Gelder, B.; de Borst, A.W.; Watson, R. (2015) The perception of emotion in body expressions. WIREs Cognitive Science 6: 149–158.
Ekman, P. (1965) Communication through nonverbal behavior: a source of information about an interpersonal relationship. In: Tomkins, S.S. & Izard, C.E. (Eds.) Affect, Cognition and Personality: Empirical Studies. Spinger, Oxford. Pp. 390–442.
Ekman, P. (1982) Emotion in the Human Face. Second Edition. Cambridge University Press, Cambridge.
Izard, C.E. (1972) Patterns of Emotion: a new analysis of anxiety and depression. Academic Press, New York.
Jennings, B.J.; Yu, Y.; Kingdom, F.A.A. (2017) The role of spatial frequency in emotional face classification. Attention, Perception & Psychophysics 79(6): 1573–1577.
Levy, J.; Foulsham, T.; Kingstone, A. (2013) Monsters are people too. Biology Letters 9(1): 20120850.
McArthur, L.Z. & Baron, R.M. (1983) Toward an ecological theory of social perception. Psychological Review 90(3): 215–238.
Meeren, H.K.; van Heijnsbergen, C.C.; de Gelder, B. (2005) Rapid perceptual integration of facial expression and emotional body language. Proceedings of the National Academy of Sciences 102: 16518–16523.
Russel, J.A. (1980) A circumplex model of affect. Journal of Personality and Social Psychology 39(6): 1161–1178.
Schmidtmann, G.; Sleiman, D.; Pollack, J.; Gold, I. (2016) Reading the mind in the blink of an eye – a novel database for facial expressions. Perception 45: 238–239.
Sekien, T. (1776) 画図百鬼夜行 [Gazu Hyakki yagyō; The Illustrated Night Parade of a Hundred Demons]. Maekawa Yahei, Japan.
About the Author
Dr. Ben Jennings is a vision scientist. His research psychophysically and electrophysiologically investigates colour and spatial vision, object recognition, emotions, and brain injury. His favourite Pokémon is Beldum.
Pocket Monsters or as they are better known, Pokémon, are playable monsters which first appeared in the 1990’s as a video game in Japan, but soon expanded worldwide. They are still very successful with numerous games, a TV series, comic books, movies, toys and collectibles, additionally to the trading card game and video games. Most recently the release of Pokémon GO, an augmented reality game for smartphones, meant that Pokémon became as popular as never before. The game launched in 2016 and almost 21 million users downloaded it in the very first week in the United States alone (Dorwald et al., 2017).
The games and TV series take place in regions inhabited by humans and Pokémon. Each Pokémon lives in a specific environment (forests, caves, deserts, mountains, fields, seas, beaches, mangroves, rivers, and marshes). The humans try to catch Pokémons with Pokéballs, a device that fits even the largest Pokémon but that is still small enough to be placed into a pocket, hence the name Pocket Monster (Whitehill et al., 2016). After Pokémon have been caught, they are put to fight against each other, just like in the real world, in which humans (unfortunately) let cockerels, crickets, or dogs fight (Marrow, 1995; Jacobs, 2011; Gibson, 2005). The origin of Pokémon goes back to the role-playing game created by Satoshi Tajiri and released by Nintendo for the Game Boy (Kent, 2001). Tajiri was not only a game developer, but like many Japanese adults, grew up catching insects as a child. He wanted to design a game so that every child in Japan could play and let their critters fight, even if they lived in areas which are too densely populated to find insects in the wild. This resulted in the 151 Pokémon in the first versions of the game (“first generation”), with each version adding more Pokémon.
Today, there are 807 Pokémon (seventh generation). Almost all are based on real organisms (mostly animals, but many plants as well), while some depict mythological creatures or objects (e.g., stones, keys). Each Pokémon belongs to one or two of the following 18 types: Normal, Fire, Fighting, Water, Flying, Grass, Poison, Electric, Ground, Psychic, Rock, Ice, Bug, Dragon, Ghost, Dark, Steel, and Fairy (Bulbapedia, 2018). All Pokémon in the game are oviparous, which means they all lay eggs; probably because the creator was fond of insects or just for practical reasons.
Certain Pokémon also evolve; however, this kind of evolution is not the same as the biological concept of evolution. In Pokémon evolution is largely synonymous to metamorphosis, such as when a caterpillar turns into a butterfly. As this is the core concept of the game, almost all Pokémon evolve, not only the insects, but also mammals, rocks, and mythological creatures. Usually, they evolve with a complete or incomplete metamorphosis: either they just grow larger, or their look differs significantly between the adult and the young stages.
Insects are the largest group of organisms on earth (Zhang, 2011). There are more than one million described species of insects, of a total of 1.8 million known organisms (Zhang, 2011). They occupy all terrestrial environments (forests, fields, under the soil surface, and in the air) and freshwater; some are even found in the ocean. Additionally, they show a wide range of morphological and behavioral adaptations. This biodiversity is not reflected in the Pokémon world. In the present Generation VII, only 77 of the 807 Pokémon are “Bug type”: about 9.5% of all Pokémon. The aim of this work is to describe the entomological diversity of Pokémon based on taxonomic criteria of the classification of real insects.
The Pokédex was the source of primary information on Pokémon (Pokémon Website, 2018). The criteria to identify insects are either based on the type (Bug type) or morphology (resembles a real insect). Afterwards, the insect Pokémon were classified to the lowest possible taxonomic level (family, genus, or species) according to their real world counterparts. This classification of the Pokémon allowed the comparison of their biological data (such as ecological or morphological traits; Bulbapedia, 2018) with the current knowledge of real insects. The information of the biology of real insects is largely based on Borror et al. (1981).
Not all Bug types are insects; many of them represent other arthropods, like spiders, while some are from other invertebrate groups (Table 1). Also, five insect Pokémon do not belong to the Bug type (e.g., Trapinch (#328) is a Ground type; Table 2). In total, insects represent only 62 of the 807 Pokémon. In comparison, the vertebrate groups are overly well-represented by birds (61), mammals (232), reptiles (57), amphibians (23), and fishes (39) (Table 3).
Eleven insect orders are represented in the Pokémon world, namely Blattodea (with 1 Pokémon), Coleoptera (11), Diptera (3), Hemiptera (7), Hymenoptera (6), Lepidoptera (22), Mantodea (4), Neuroptera (3), Odonata (2), Orthoptera (2), Phasmatodea (1). They are listed below in systematic order.
Families: Libellulidae and Aeshnidae
Genera: Erythrodiplax and Anax
Yanma (#193) evolves to Yanmega (#469).
Yanma is a large, red dragonfly Pokémon. Like all dragonflies and damselflies, it lives near the water and hunts other insects for food. Yanma is territorial and prefers wooded and swampy areas. Based on its appearance, it belongs to the dragonfly family Libellulidae, and further to the genus Erythrodiplax Brauer, 1868.
Yanmega on the other hand is a large, dark green Pokémon. It is actually a different real-world species. Not only the colors are different, but also the morphology, like the appendages on the tip of the tail. Based on this, it belongs to the dragonfly family Aeshnidae, and to the genus Anax Leach, 1815. One could argue that it is based on Meganeura Martynov, 1932, a very large (wingspan up to 70 cm) but extinct dragonfly genus from the Carboniferous Period. However, the size alone should not be the indicator to classify the species, as many insectoid species are larger in the Pokémon world compared to the real world.
Scyther (#123) evolves to Scizor (#212, incl. Mega-Scizor).
Scyther is a bipedal, insectoid Pokémon. It is green with cream joints between its three body segments, one pair of wings and two large, white scythes as forearms. Scyther camouflages itself by its green color. Based on its appearance, it is classified as a praying mantis (or possible a mantidfly).
Scizor is also a bipedal, insectoid Pokémon. It is primarily red with grey, retractable forewings. Scizor’s arms end in large, round pincers. It appears to be based on a praying mantis, maybe with some references to flying red ants and wasp-mimicking mantidflies.
Although Scizor evolves from Scyther, they are very different and would actually be two different real-world species. Not only are the colors different, but also the morphology: the arms end in either scythes or pincers; Scyther has one pair of wings, Scizor has two.
Fomantis (#753) evolves to Lurantis (#754).
Fomantis is a plant-like and, at the same time, an insect-like Pokémon. Its main body is pink, with green hair, green tufts on the head, and green leaves as a collar. Fomantis is somewhat bipedal and is likely based on the orchid mantis Hymenopus coronatus Olivier, 1792 (Fig. 1), which is known for being able to mimic the orchid flower, along with the orchid itself.
Lurantis is also plant- and insect-like. It is pink, white, and green. Lurantis looks and smells like a flower, to attract and then attack foes (and prey). It also disguises itself as a Bug Pokémon for self-defense. Lurantis is likely based on the orchid mantis as well as the orchid flower itself, as it is impossible to say where the flower ends and the insect starts. Orchid mantises mimic parts of a flower, by making their legs look like flower petals. Well camouflaged, they can wait for their prey, which will visit the flower for nectar.
Pheromosa is a bipedal anthropomorphic Pokémon. It has a rather slender build and is mostly white. Pheromosa originates from the Ultra Desert dimension in Ultra Space. Pheromosa is based on generic cockroaches just after they have molted (Fig. 2); during this stage, the animals are pale and vulnerable until their exoskeleton hardens and darken.
Kricketot (#401) evolves to Kricketune (#402).
Kricketot is a bipedal, bug-like Pokémon. It has a red body with some black and white markings. By shaking its head and rubbing its antennae together, it can create a sound that it uses to communicate. Based on its appearance, it is a cricket.
Kricketune is also a bipedal Pokémon with an insectoid appearance, also primarily red with some black and tan colored markings. It can produce sound by rubbing its arms on the abdomen. Kricketune appears to be based on crickets due to their sound-producing ability, but it somewhat resembles a violin beetle.
Both Kricketot and Kricketune are depicted with only 4 limbs, whereas insects are largely defined by having exactly six legs.
Families: Gerridae and Fulgoridae
Surskit (#283) evolves to Masquerain (#284).
Surskit is a blue insectoid Pokémon with some pink markings. It produces some sort of syrup, which is exuded as a defense mechanism or to attract prey. This Pokémon can also secrete oil from the tips of its feet, which enables it to walk on water as though skating. Surskit usually inhabits ponds, rivers, and similar wetlands, where it feeds on microscopic, aquatic organisms. This Pokémon is based on water striders. However, a water strider does not ooze syrup and neither does it need oil to walk on water; it can walk on water due to the natural surface tension.
Masquerain is a light blue Pokémon with two pairs of wings. On either side of its head is a large antenna that resembles an angry eye. These eyespots are used by many real-life moths and lantern-flies to confuse and intimidate would-be predators. Masquerain is in fact based on a lantern-fly.
Both “species”, water striders and lantern-flies, are only distantly related, belonging to two different families within the “true bugs” (Hemiptera).
Nincada (#290) evolves to Ninjask (#291) and then to Shedinja (#292).
Nincada is a small, whitish, insectoid Pokémon. The claws are used to carve the roots of tree and absorb water and nutrients. Nincada builds underground nests by the roots of trees. It is based on a cicada nymph, which lives underneath the soil surface. However, a cicada nymph usually does not have fully developed wings. Instead, they have short wing stubs which eventually will become fully functional wings – as usual amongst hemimetabolous insects.
Ninjask is a small, cicada-like Pokémon with two pairs of wings. Its body is mostly black with some yellow and grey markings. Ninjask is a very fast Pokémon and it can seem invisible due to its high speed. It is based on an adult cicada, with the colors somewhat resembling Neotibicen dorsatus (Say, 1825) (Fig. 3).
Shedinja is a brown and grey insectoid Pokémon. A hole between its wings reveals that its body is completely hollow and dark, as it possesses no internal organs. It is based on the shed husk (exuvia) that cicadas and other hemimetabolous insects leave behind when they molt.
Paras (#046) evolves to Parasect (#047).
Paras is an orange insectoid Pokémon with an ovoid body. On the top it has two little red and yellow mushrooms known as tōchūkasō. The mushrooms can be removed at any time, and grow from spores that are doused on this Pokémon’s back at its birth by the mushroom on its mother’s back. Tōchūkasō is an endoparasitoid that replaces the host tissue and can affect the behavior of its insect host. The base insect is based on a cicada nymph. The real-world tōchūkasō live on hepialid caterpillars in Tibet. However, there are many more species of entomopathogenic fungi in the world, most notable the genus Cordyceps (L.) Fr. (1818).
Parasect is an orange, insectoid Pokémon that has been completely overtaken by the tōchūkasō mushroom. The adult insect has been drained of nutrients and is now under the control of the fully-grown tōchūkasō. Parasect can thrive in dank forests with a suitable amount of humidity for growing fungi. The base insect is a deformed version of what is probably a cicada nymph, the parasitic mushroom having caused a form of neoteny, when the adults look like a juvenile form.
Trapinch (#328) evolves to Vibrava (#329) and then to Flygon (#330).
Trapinch is an orange, insectoid Pokémon. This Pokémon lives in arid deserts, where it builds its nest in a bowl-shaped pit dug in sand. It sits in its nest and waits for prey to stumble inside. Once inside, the prey cannot climb back out. It is based on the larval stage of the antlion, which lives in conical sandy pits before maturing into winged adults.
Vibrava is a dragonfly-like Pokémon. Vibrava’s wings are not fully developed, so it is unable to fly very far. However, it is able to create vibrations and ultrasonic waves with its wings, causing its prey to faint. Vibrava is a saprotroph – it spits stomach acid to melt its prey before consumption. Vibrava is based on the adult stage of an antlion. Adult antlions and dragonflies look from a distance quite similar and are therefore often mistaken for each other.
Flygon is a desert-dwelling insectoid dragon with a green body and one pair of wings. Its wings make a “singing” sound when they are flapped. It uses this unique ability to attract prey, stranding them before it attacks. It is based on the winged, adult stage of the antlion.
Pinsir (#127, incl. Mega-Pinsir).
Pinsir is a bipedal beetle-like Pokémon with a brown body and a large pair of grey, spiky pincers on top of its head. Pinsir is based on a stag beetle.
Grubbin (#736) evolves to Charjabug (#737) and then to Vikavolt (#738).
Grubbin is a small insectoid Pokémon. It has a white body with three nubs on either side resembling simple legs. Grubbin typically lives underground. It uses its jaw as a weapon, a tool for burrowing, and for extracting sap from trees. Grubbin appears to be based on a larval beetle, also known as “grubs”.
Charjabug is a small cubic Pokémon resembling an insect-like battery. Its body consists of three square segments with two brown stubs on each side. It generates and stores electricity in its body by digesting food. This energy is stored in an electric sac. Charjabug appears to be based on a cocooned bug and a battery. It may also be based on the denkimushi (Monema flavescens Walker, 1855), a caterpillar in Japan that, when touched, can give a sting that is said to feel like an electric shock (Fig. 4).
Vikavolt is a beetle-like Pokémon with a large pair of mandibles. It produces electricity with an organ in its abdomen, and fires powerful electric beams from its huge jaws. Vikavolt appears to be based on a stag beetle. Its straight, scissor-like mandibles resemble those of Lucanus hayashii Nagai, 2000.
Ledyba (#165) evolves to Ledian (#166).
Ledyba is a red ladybird-like Pokémon with five black spots on its back. Female Ledyba have shorter antennae than male Ledyba. Ledyba is a very social Pokémon, e.g. in the winter they gather together to keep each other warm. Ledyba is probably based on the five-point ladybird Coccinella quinquepunctata Linnaeus, 1758 due to its color and/or on the harlequin ladybird Harmonia axyridis (Pallas, 1773), which clusters together in the winter.
Ledian is a large red bipedal ladybird-like Pokémon. Female Ledians’ antennae are shorter than the males’. Ledian sleeps in forests during daytime inside a big leaf.
Heracross (#214, incl. Mega-Heracross).
Heracross is a bipedal beetle-like Pokémon with a blue exoskeleton. The prolonged horn on its forehead ends in a cross-shaped (males) or heart-shape (females) structure. Heracross is most likely based on the Japanese rhinoceros beetle Allomyrina dichotoma Linneaus, 1771 (Fig. 5).
Volbeat (#313) and Illumise (#314).
Volbeat is a bipedal firefly-like Pokémon. Its body is black with some blue, yellow, and red portions. It has a spherical yellow tail, which glows to communicate and draws geometric patterns in the sky while in a swarm. This is a male only Pokémon “species”; Illumise is its female counterpart. Volbeat lives in forests near clean ponds and is attracted by the sweet aroma given off by Illumise. It is based on a firefly like its counterpart Illumise. Its appearance may be based on a greaser, a subculture from the 1950’s.
Illumise is a bipedal firefly-like Pokémon. It is black and blue with some yellow markings. This is a female only Pokémon “species”; Volbeat is its male counterpart. It is a nocturnal Pokémon that lives in forests. Illumise does not seem to share its coloring with any particular species. Illumise may be based on flappers, a 1920’s women’s style. Its mating behavior only slightly resembles the behavior of real-world fireflies, in which females use light signals to attract mates.
Karrablast (#588) evolves to Escavalier (#589).
Karrablast is a round bipedal Pokémon with a yellow and blue body. When it senses danger, it spews an acidic liquid from its mouth. It targets another Pokémon, Shelmet, so it can evolve. It resides in forests and fields, and it often hides in trees or grass if threatened. Karrablast may be based on a Japanese snail-eating beetle due to its preference for attacking Shelmet, a snail-like Pokémon.
Escavalier is an insectoid Pokémon wearing a knight’s helmet. Its tough armor protects its entire body. It flies around at high speed, jabbing foes with its lances. Escavalier is probably based on the Drilus Olivier, 1790 genus, with references to a jousting knight. Drilus larvae are known for eating snails and stealing their shells, explaining why it attacks Shelmet and takes its shell to evolve into Karrablast.
Weedle (#013) evolves to Kakuna (#014) and then to Beedrill (#015, incl. Mega-Bedrill).
Weedle is a small larval Pokémon with a body ranging in color from yellow to reddish-brown. It has a conical venomous stinger on its head and a barbed one on its tail to fend off enemies. Weedle can be found in forests and usually hides in grass, bushes, and under the leaves it eats. Weedle appears to be based on the larva of a wasp or hornet, although these real-world larvae usually don’t have defense strategies. The only larvae which feed directly off leaves are those of sawflies.
Kakuna is a yellow cocoon-like Pokémon. Kakuna remains virtually immobile and waits for its “evolution” to happen, often hanging from tree branches by long strands of silk. Although Kakuna is the pupa stage of a Hymenoptera, it showcases a silky cocoon, a feature usually found in Lepidoptera and only some Hymenoptera, like sawflies.
Beedrill is a bipedal, wasp-like Pokémon. Its forelegs are tipped with long, conical stingers. It stands on its other two legs, which are long, segmented, and insectoid in shape. Beedrill has two pairs of rounded, veined wings, and another stinger on its yellow-and-black striped abdomen. By its color pattern, Beedrill looks like a vespid wasp, but due to the previous stages of this Pokémon species, it must be based on Tenthredo scrophulariae Linneaus, 1758, the figwort sawfly.
Combee (#415) evolves to Vespiquen (#416, female).
Combee is a small insectoid Pokémon that resembles three social bees inside three hexagonal pieces of honeycomb stuck together; the top two have wings. Female Combee have a red spot on the forehead. Male Combee are not known to evolve into or from any other Pokémon. The sex ratio of Combee is 87.5% male and 12.5% female. Combee can fly with its two wings as long as the top two bees coordinate their flapping. They gather honey, sleep, or protect the queen. Combee is based on a mix of bees and their larvae living in honeycombs. (Bees arrange their honeycombs in a vertical manner, whereas wasps arrange them horizontally.)
In the hive of the real-world honey bee (Apis mellifera Linneaus, 1758), there is usually one queen bee and up to 40.000 female workers. So, the sex ratio of Combee does not reflect the ratio of female (workers) and male (drones) honey bees, but of the reproductive bees, the drones and the fertile queens. The larger number of drones is needed, since each queen will often mate with 10–15 males before she starts a new hive. Usually, drones can make up to 5% of the bees in a hive.
Vespiquen is a bipedal bee-like Pokémon with a yellow and black striped abdomen resembling an elegant ballroom gown. Underneath the expansive abdomen are honeycomb-like cells that serve as a nest for baby Combee. Vespiquen is a female-only Pokémon “species”. Vespiquen is the queen of a Combee hive, controlling it and protecting it, as well as giving birth to young Combee. The horizontal honeycombs hints that this “species” is a wasp rather than a bee.
Durant is an ant-like Pokémon with a grey body and six black legs. It is territorial, lives in colonies and digs underground mazes. Durant grows steel armor to protect itself from predators. Durant is based on an ant, possibly the Argentine ant (Linepithema humile Mayr, 1868), due to the jaw and their invasive behavior.
Caterpie (#010) evolves to Metapod (#011) and then to Butterfree (#012).
Caterpie is a green caterpillar-like Pokémon. It has yellow ring-shaped markings down the sides of its body and bright red “antenna” (osmeterium) on its head, which releases a foul odor to repel predators. The appearance of Caterpie helps to startle predators; Caterpie is probably based on Papilio xuthus Linnaeus, 1767, the Asian swallowtail (Fig. 6). The osmeterium is a unique feature of swallowtails. Caterpie will shed its skin many times before finally cocooning itself in thick silk. Its primary diet are plants.
Metapod is a green chrysalis Pokémon. Its crescent shape is based upon a Swallowtail chrysalis with a large nose-like protrusion and side protrusions resembling a Polydamas Swallowtail or Pipevine Swallowtail chrysalis (genus Battus Scopoli, 1777).
Butterfree is a butterfly Pokémon with a purple body and large, white wings, somewhat resembling a black-veined white Aporia crataegi (Linneaus, 1758). Although it is supposed to be a butterfly, it lacks the proboscis, which is typical of Lepidoptera, and presents teeth instead. Additionally, the body does not consist of the typical three segments of insects. Therefore, each stage seems to be based on a different species.
Families: Geometridae and Arctiidae
Venonat (#048) evolves to Venomoth (#049).
Venonat has a round body covered in purple fur, which can release poison. It feeds on small insects, the only Lepidoptera caterpillar which is known to feed on prey instead of leaves belong the genus Eupethecia Grote, 1882 (Geometridae). However, Venonat does not resemble a caterpillar in general body shape or numbers of legs.
Venomoth is a moth-like Pokémon with a light purple body and interestingly two small mandibles. It has two pairs of wings, which are covered in dust-like, purple scales, although the color varies depending on their toxic capability. Dark scales are poisonous, while lighter scales can cause paralysis. These scales are released when Venomoth flutters its wings. The general appearance resembles species belonging to the Actiidae.
There is no cocoon stage for this species it is doubtful whether both stages were based on the same real-life species.
Scatterbug (#664) evolves to Spewpa (#665) and then to Vivillon (#666).
Scatterbug is a small caterpillar Pokémon with a grey body. If threatened by a bird Pokémon, it can spew a powder that paralyzes on contact. Similarly, the large white butterfly Pieris brassicae (Linneaus, 1758) is known to throw up a fluid of semi-digested cabbage, which contains compounds that smell and taste unpleasant to predators, such as birds.
Spewpa is a small insectoid Pokémon with a grey body covered by white furry material. In order to defend itself, Spewpa will bristle its “fur” to threaten predators or spray powder at them. Spewpa is based on a generic pupa of a moth or butterfly, probably a silkworm cocoon.
Vivillon is a butterfly-like Pokémon with wings that come in a large variety of patterns, depending in which climate it lives or rather, in which real-world region the player is. There is a total of 20 patterns known. It would be interesting to know whether they evolved due to allopatric speciation or if it is a case of mimicry.
Pineco (#204) evolves to Forretress (#205).
Pineco is a pine cone-like Pokémon without visible limbs. It is based on a bagworm, the caterpillar stage of psychid Lepidoptera. Bagworms cover themselves with a case (the bag) made of surrounding material. This Pokémon uses tree bark and thus resembles a pine cone.
Forretress is a large spherical Pokémon, also without any visible limbs. It lives in forests, attaching itself immovably to tree trunks. Forretrees is also based on a bagworm.
Different bagworm species are adapted to their environment, to the plants they eat, and to the materials available for producing their case. Therefore, Pineco and Forretress are actually based on two different species, as they both are caterpillars. There is no adult stage for this Pokémon.
Burmy (#412) evolves to Wormadam (#413, female) or Mothim (#414, male).
Burmy is a small pupa-shaped Pokémon with a black body and six stubby legs. It is based on a bagworm pupa, which will metamorphose into a winged moth if male, or wingless moth if female. Burmy can change its “cloak” (case) depending on the environment it last battled.
Wormadam is a black bagworm-like Pokémon with a cloak of leaves, sand, or building insulation. Its cloak depends on Burmy’s cloak when it evolved, and so does it type (Grass, Ground or Steel). It is a female-only “species”, with Mothim as its male counterpart. Female psychid moth either don’t have wings at all or have only small wing stubs that don’t develop fully.
Mothim is a moth-like Pokémon with two pairs of legs and two pairs of wings, one larger than the other. Mothim is a nomadic nocturnal Pokémon, searching for honey and nectar. Instead of gathering honey on its own, it raids the hives of Combee. It is a male-only “species”, with Wormadam as its female counterpart.
Wurmple (#265) evolves to Silcoon (#266) and then to Beautifly (#267).
Wurmple is a small caterpillar-like Pokémon with a mostly red body and many spikes on the top of its body. It can spit a white silk that turns gooey when exposed to air. Spikes or hairy appendages are common amongst nymphalid caterpillars. Also, it has five pairs of legs, whereas insects are known to have only three pairs of legs. However, many lepidopteran caterpillars have additionally “prolegs” (small fleshy stub-like structures) to help them move.
Silcoon is a cocoon-like Pokémon which is completely covered by white silk. Silcoon also uses the silk to attach itself to tree branches. Nymphalid cocoons are usually not woolly or hairy, but smooth.
Beautifly is a butterfly-like Pokémon with two pairs of wings. Beautifly has a long and curled black proboscis that it uses to drain body fluids from its prey. In the real world, Lepidoptera usually drink the nectar of flowers. One of the few exceptions are the species of the genus Calyptra Ochsenheimer, 1816, which pierce skin of animals and drink blood.
Wurmple (#265) evolves to Cascoon (#268) and then to Dustox (#269).
The caterpillar stage of this species is morphologically identical to the caterpillar stage of the “species” above: Wurmple. It appears that Wurmple can evolve in two forms: due to mimicry, sympatric speciation or are there morphological or biological characters, which have not been notices yet?
Cascoon is a round cocoon-like Pokémon covered in purple silk. Saturniid cocoons are usually covered in silk.
Dustox is a moth-like Pokémon. It has a purple body, two pairs of tattered green wings, and – just like Beautifly – two pairs of legs. Dustox is nocturnal and is instinctively drawn to light. Clearly, this is a moth. Some of the markings on its wings resemble typical markings of noctuid moths, but the big “fake eye” is typical of saturniids.
Larvesta (#636) evolves to Volcarona (#367).
Larvesta is a fuzzy caterpillar-like Pokémon. It has five red horns on the sides of its head, which it can use to spit fire as a defensive tactic to deter predators. Larvesta is based on a saturniid caterpillar.
Volcarona is a large moth-like Pokémon with four small feet and three pairs of wings. It releases fiery scales from its wings. Just like Larvesta, Volcarona is based on a saturniid moth, likely the Atlas moth Attacus atlas (Linneaus, 1758).
Cutiefly (#742) evolves to Ribombee (#743).
Cutiefly is a tiny Pokémon with large wings. Cutiefly appears to be based on the bee fly, specifically the species Anastoechus nitidulus (Fabricius, 1794) (Fig. 7).
Ribombee is a tiny insectoid Pokémon with a large head, slightly smaller body, and thin arms and legs. It is covered in fluffy yellow hair. Two wings nearly as large as its body sprout from its back. The wings are clear with three brown loop designs near the base. Its four thin limbs have bulbous hands or feet. Ribombee uses its fluffy hair to hold the pollen it collects from flowers. It is based on a bee fly.
Buzzwole is a bipedal anthropomorphic Pokémon. It has four legs and two pairs of orange translucent wings. It uses its proboscis to stab and then drink “energy” off its enemies/prey. Buzzwole originates from the Ultra Desert dimension in Ultra Space. It is based on a mosquito and may specifically derive inspiration from Aedes albopictus (Skuse, 1894), which is an invasive species worldwide.
Mixed Orders: Lepidoptera and Phasmatodea
Families: Tortricidae, Hesperiidae, and Phylliidae
Sewaddle (#540) evolves to Swadloon (#541) and then to Leavanny (#542).
Sewaddle is a caterpillar-like Pokémon with a green body with three pairs of legs. It makes leafy “clothes” using chewed-up leaves and a thread-like substance it produces from its mouth. The leafy hood helps Sewaddle to hide from enemies. Sewaddle appears to be based on the caterpillar of the silver-spotted skipper Epargyreus clarus (Cramer, 1775), which produce silk and fold leaves over themselves for shelter (Fig. 8).
Swadloon is a round yellow Pokémon inside of a cloak of leaves. It lives on the forest ground and feeds on fallen leaves. Swadloon appears to be based on the chrysalis of Epargyreus clarus. Epargyreus clarus fold leaves over themselves for shelter as they age and, when cocooning, eventually use silk to stick the leaves together and form its chrysalis.
Leavanny is a bipedal, insectoid Pokémon with a yellow and green body with leaf-like limbs. It lives in forests and uses its cutters and sticky silk it produces to create leafy “clothing”. It also warms its eggs with fermenting fallen leaves. Leavanny has the features of several insects. Primarily it appears to be a bipedal leaf-insect (Phylliidae). Its general body structure is also similar to that of Choeradodis Serville, 1831 mantises, which also have laterally expanded thoraxes and abdomens.
Only 11 insect orders (out of 30) are represented in the Pokémon world. Possible more, as differentiation of insect Pokémon and non-insect Pokémon are sometimes difficult. The main reason is, that many insect Pokémon are not depicted as a typical insect with its segmented body, the six legs, and two pairs of wings. Many are depicted as bipedal (e.g., #401 Kricketot) or even in an anthropomorphic way (e.g., #795 Pheromosa). Also, insectoid Pokémon typically have only four limbs (instead of six). Many insectoid Pokémon also have fewer wings than insects (except for #637 Volcarona, which has more). Therefore, the definition of what is an insect Pokémon is debatable.
One clue is to look at the types each Pokémon belongs to. However, from the circa 80 Bug-type Pokémon, only about 60 are insects. The others belong to other arthropods groups, like Chelicerata, Crustacea, and Myriapoda. This is not surprising, as often creepy crawlies (basically everything that is small with legs) are all addressed as “bugs”. In fact, only member of the insect order Hemiptera are called “true bugs”.
Interestingly, Prado & Almeida (2017) have included Pokémon on their insect list, which are doubtful: #251 Celebi, #247 Pupitar, and #206 Dunsparce. None of them are considered insects here. Celebi may resemble a bipedal somewhat anthropomorphic insectoid, but nothing of the lifestyle or beyond the vague appearance gives a clue to an insect. Similarly, #247 Pupitar, might look like a pupa of an insect. However, both its “larval” stage (#256 Larvitar) and its final stage (#248 Tyranitar) resemble a dinosaur or some sort of dragon. Only the hint of “pupa” in its name, links Pupitar to an insect. Lastly, #206 Dunsparce was classified as a Hymenoptera by Prado & Almeida (2017). Is may look somewhat like an insect, even showing two pairs of wings (and no legs at all). Dunsparce, however, is based on a mythical “snake-like animal” called Tsuchinoko, also known as “bachi hebi” (or “bee snake”). Finally, Prado & Almeida (2017) have classified #212 Scizor as “unknown”, but here it is treated as a praying mantis (Mantodea). Similarly, those authors have classified #284 Masquerain as a Lepidoptera, but here we treat is as a true bug (Hemiptera).
Lastly, #649 Genesect resembles somewhat an ant covered by steel. However, according to the Pokédex (Pokémon Website, 2018), it is a man-made machine.
Compared to the vertebrates (birds, mammals, reptiles, amphibians, and fishes), many more insects live on earth (66,000 described species to about 1 million, respectively; Zhang, 2011). This ratio is, however, not represented in the Pokémon world (Table 3), most likely due to the fact that the majority of people prefer (cute and cuddly) furry animals over creepy insects, even though butterflies and dragonflies are regarded as beautiful.
Borror, D.J.; DeLong, D.M.; Triplehorn, C.A. (1981) An Introduction to the Study of Insects. Saunders College, Philadelphia.
Bulbapedia (2018) The community driven Pokémon encyclopedia. Available from: http://bulbaped ia.bulbagarden.net/ (Date of access: 10/Sep/ 2018).
Dorward, L.J.; Mittermeier, J.C.; Sandbrook, C.; Spooner, F. (2017) Pokémon GO: benefits, costs, and lessons for the conservation movement. Conservation Letters 10(1): 160–165.
Gibson, H. (2005) Detailed Discussion of Dog Fighting. Michigan State University, East Lansing.
Kent, S.L. (2001) The Ultimate History of Video Games. Crown Publishing Group, New York.
Morrow, L. (1995) History they don’t teach you: a tradition of cockfighting. White River Valley Historical Quarterly 35(2): 5–15.
Official Pokémon Website, The. (2018) The Official Pokémon Website. Available from: http://poke mon.com/ (Date of access: 10/Sep/2018).
Prado, A.W. & Almeida, T.F.A. (2017) Arthropod diversity in Pokémon. Journal of Geek Studies 4(2): 41–52.
Whitehill, S.; Neves, L.; Fang, K.; Silvestri, C. (2016) Pokémon: Visual Companion. Pokémon Company International / Dorling Kindersley, London.
Zhang, Z.-Q. (2011) Animal biodiversity: an outline of higher-level classification and survey of taxonomic richness. Zootaxa 3703: 1–82.
I am grateful to Seth Ausubel (https://www. flickr.com/photos/96697202@N07/collections) for kindly granting permission to use his photograph of Epargyreus clarus on this article. I would also like to thank Miles Zhang for valuable comments on an earlier version of the manuscript.
ABOUT THE AUTHOR
Dr. Rebecca Kittel is an entomologist working on parasitoid wasps. She is interested in all sorts of interactions of insects with human beings, regardless of whether they are real-life insects or purely fictional.
 Not all insects have two pairs of wings, though. For instance, the Diptera (flies) have only one, while the Siphonaptera (fleas) have none.
As an avid consumer of Japanese video games during my early teens, particularly of the RPG sort, I could not help but notice that some monsters would pop up in several games and typically had a pretty standard depiction. I have always been interested in mythology and could naturally identify the usual chimeras, griffins, phoenixes, and gorgons.
However, these monsters shared their screen time with more unusual ones (or unusual to me at least) from Japanese myths and folklore. Maybe expectedly, I started to read about Japanese myths and to learn about kappa, tengu and many others. Still, one monster, in particular, was suspiciously absent from the books: a sort of statue-like creature with large round eyes (Fig. 1). I did not know its actual name and could not find information about it anywhere.
Then, I forgot all about this monster when I switched my geek focus to tabletop RPGs and my gaming preferences to Western hits (Bioware RPGs, Gears of War, etc.). This lasted until some years ago when I played Persona 4 and Pokémon: Alpha Sapphire for the first time (I had skipped Pokémon’s Gen III back in the day); there and then, I re-encountered that weird statue-like creature (Fig. 2).
Even so, it was not until a recent visit to the British Museum that my interest was reignited. In their Japanese exhibition, I discovered that this creature was not a mythological monster after all — it was nothing like a tengu or a kappa! The damn thing was a prehistoric clay figurine (Fig. 3). As a category, these figurines are called “dogū”.
Needless to say, I began searching for books and scholarly articles about dogū. Sadly, most of the literature on them (and prehistoric Japan in general) is in Japanese, which I cannot read and do not trust Google to translate it for me. Nevertheless, I wanted to report what I could find, just in case these figurines have captured the imagination of someone else out there (maybe someone like you, dear reader). So please keep in mind that my report here is based on the somewhat scarce literature available in English and thus it may lack some information and/or be overly simplified in some aspects.
Before we start, however, I need to briefly explain how Japanese prehistory is divided. So let’s get down to it.
Japanese prehistory can be broadly divided into two large periods: the Paleolithic and what may be informally called “Ancient Japan” (Table 1). The latter is a mixture of the usual Mesolithic, Neolithic and Bronze Age that has defied classification by archaeologists using this standard Western periodization (Imamura, 1996). This span of time contains three periods: the Jōmon, the Yayoi, and the Kofun. Here we are interested only in the first one, the Jōmon period.
Taken literally, Jōmon means “cord-marked”. This refers to the usage of cords to create decorative patterns on ceramics (Fig. 4), which was achieved by simply pressing a cord on the clay prior to firing (Kaner, 2009).
During the Jōmon period, Japan was covered by rich temperate forests (Imamura, 1996). This allowed people to live as hunter-gatherers, although there were phases (maybe seasonal) of sedentism, with some settlements growing quite large and possibly housing a few hundred inhabitants (Imamura, 1996; Henshall, 2004). There is also evidence of slash-and-burn agriculture and limited domestication of plant species, accompanied by skillful management of resources (Imamura, 1996; Habu, 2004). Furthermore, a good portion of the Jōmon people lived close to the coast, exploring marine resources (Henshall, 2004).
The Jōmon period was not, however, a single homogenous thing across all Japan. There was regional variation in habits and material culture, which changed at different paces throughout the country (Henshall, 2004). Furthermore, people from the continent migrated into Japan and added their share of knowledge, culture and genes to the mixture (Imamura, 1996). The Jōmon period ended with the start of rice cultivation and metallurgy.
One important social aspect that gained strength during the Jōmon was how people dealt with the supernatural. Artifacts (Fig. 5), burial practices, and stone circles (Fig. 6) all indicate that religion and ritual were steadily developing throughout the period (Kaner, 2011). One type such artifacts was, of course, the dogū.
Dogū are ceramic figures produced during the Jōmon period. The earliest dogū dates back to the Incipient Jōmon (Table 2) and they remained restricted in numbers during the Initial and Early Jōmon (Habu, 2004). However, from the Middle Jōmon onwards, their manufacture thrived and their design became more elaborate (Kaner, 2009).
Most of the dogū are clearly female (some of them supposedly pregnant; Fig. 7), so some scholars believe they are representations of an earth-goddess. They claim that this mother-goddess worship is common in agricultural societies, but then again, agriculture was only incipient during the Jōmon period. Other scholars take into consideration the prominence of secondary sex characteristics and hypothesize that the dogū are just general fertility symbols, related to fertility rituals and magical protection during dangerous events such as childbirth. This latter option seems apparently more likely, as similar symbols are known from pretty much everywhere.
Nevertheless, considering that figurines such as these have only one function is careless, to say the least (Soffer et al., 2000). As such, other interpretations have appeared in the last decades. For instance, some authors link the increase in the production of dogū from the Middle Jōmon onwards to an increase of agricultural practices and the role of women in this subsistence shift (Togawa, 2003).
The actual functions of dogū remain unknown, but the constant debate makes archaeologists revisit old ideas, propose new ones, and slowly fine-tune our knowledge.
There are several types of dogū, roughly classified by how they look. Because of that, they have some really amusing names (Habu, 2004): heart-shaped dogū (Fig. 8), sitting dogū, mountain-shaped-head dogū, goggle-eyed (or slit-goggle) dogū (Figs. 3, 9), horned-owl dogū.
It is still unclear if these different categories of dogū had distinct purposes or functions. Furthermore, dogū came in several sizes, from palm-sized figurines to large ones more than 30 cm high (Togawa, 2003; Kaner, 2009). As such, it is likely that they had different functions, ranging from personal belongings to probably community-wide ceremonial artifacts (Togawa, 2003).
Today, people can see all sorts of dogū in museum exhibitions around the world, like in the Tokyo National Museum and the British Museum. But they are not merely relics of an ancient past – Japanese people certainly have not forgotten them. For instance, there are some conspicuous monuments in Japan commemorating the most popular type of dogū, the goggle-eyed dogū (or shakōki-dogū).
Two of such monuments can be found in the city of Tsugaru, in Aomori prefecture. The Kamegaoka Site, an archaeological site dating from the Final Jōmon (1,000–300 BCE), is located there. This site is important because it is the place where the most textbook-famous dogū (a goggle-eyed one with a broken leg; Fig. 9) was found back in 1887 (Tsugaru City Board of Education, 2018). One of the monuments is a simple statue (Fig. 10), as could be expected, but the city’s railway station (Fig. 11) is something else entirely!
Box 1. Pseudoarchaeology
Unfortunately, the dogū (especially the goggle-eyed) became victims of human stupidity, just as several other archaeological icons (the pyramids, the Antikythera mechanism, the Nazca lines, etc.). That is, they were linked to alien activity by people who abhor scientific research and methodology and who prefer to make up their own wild stories about reality. Their “explanation” is that the goggle-eyed dogū resembles a person in a space suit. And no, I will not give the reference to their original “works” — these people should not be given the satisfaction of an actual citation!
Given the cultural importance of the dogū in Japan and the increasing influence of television, mangas and video games, it was expected that these clay figures would make their way into pop culture. This is especially true for the fan-favorite type, the goggle-eyed dogū (Rousmaniere, 2009).
The obvious examples, as I mentioned above, come from video games, especially RPGs such as the ever-present Final Fantasy (Square Enix, 1987–present) and Dragon Quest series. The dogū are featured in various games, often just as meaningless enemies in random dungeons. Thus, I will not bore you to death with an extensive list of all dogū appearances. Instead, I will point out just a few examples that I find more meaningful.
One of them is the Pokémon Claydol (Fig. 2), which does not have the most creative name around. It is a Ground / Psychic type and most Pokédex entries on the series point out that it is a clay statue made by ancient people (Bulbapedia, 2018). The entries in Pokémon Sapphire (2002), Black/White 2 (2012) and Alpha Sapphire (2014) date them from 20,000 years ago, which, as we have seen above (Table 1), is a clear exaggeration for the late parts of the Jōmon period. However, the Pokédex entry in Pokémon Ultra Moon (2017) is much more problematic; it reads: “The ancient people who made it apparently modeled it after something that descended from the sky.” Pokémon, of course, is not known for its scientific rigor (Tomotani, 2014; Mendes et al., 2017), but spreading ridiculous alien stories is irresponsible, to say the least (see also Box 1).
Another interesting appearance of the goggle-eyed dogū is in the Shin Megami Tensei series (henceforth SMT; Atlus/Sega, 1987–present), which includes the Persona sub-series. These games allow players to summon mythological monsters (and deities) from virtually all cultures around the world. Since it is a Japanese game, it focuses heavily on Japanese creatures. The goggle-eyed dogū from SMT is called Arahabaki (Fig. 12).
The entries about Arahabaki in the SMT games’ lore describe it as a god (Megami Tensei Wiki, 2018), which we have already established is the less likely hypothesis. The game also refers to it as “he/him” (at least in the English translation), while clearly depicting it with a female body, like the original clay figurines. SMT uses myths as a basis for its setting and story, and infuse them with fiction, so it is hard to tell if their information came from somewhere or if they just made it up to fill a narrative purpose. In any event, their description of the goggle-eyed dogū is off the mark.
Last but not least, there’s Ōkami (Capcom, 2006). The game is set in classical Japan and mixes lots of Japanese myths and folklore. In Ōkami, the goggle-eyed dogū (Fig. 13) is one among many demons that the player faces. The demon’s entry in the game’s bestiary (Okami Wiki, 2018) handles the matter much better than Pokémon: “Of all the odd clay figures in this land, the Dogu is the strangest. Fascinated people have speculated that they originated on the moon.” Thus, the game makes clear that the whole alien thing is just a story made up by some crazy folk.
Dogū are also featured in several mangas (e.g., Doraemon), typically as the focus of one or a handful of chapters. However, one title features them prominently: it is called “Dogū Family” (translation) and was printed in the late 1980’s and early 90’s. The story focused on the everyday life of a family of goggle-eyed dogū in modern Japan. Unfortunately, I could not find the actual manga to read.
Dogū also appear in Japanese products and TV commercials, and there is even one TV show about them: The Ancient Dogoo Girl (“Kodai Shōjo Doguchan”; Fig. 14) and its sequel The Ancient Dogoo Girls (“Kodai Shōjotai Dogūn Faibu”). The series aired on MBS (Mainichi Broadcasting System) from 2009 to 2010.
The series’ plot is very basic Japanese stuff: Makoto, a hikikomori, finds a weird breastplate buried in the woods, touches it, and awakens a girl named Dogu-chan. She is a yōkai hunter from the Jōmon period and ends up living with Makoto. Dogu-chan has a familiar/assistant named Dokigoro (Fig. 15), which is a sentient goggle-eyed dogū that transforms into magical (bikini) armor for its master. The sequel had another five girls wearing armors based on other types of dogū.
The Ancient Dogoo Girl is a very weird and rather embarrassing show, even by Japan standards, as it involves a lot of breasts-based magic. I just skimmed through the first episode to write these paragraphs and already regret it. So if you are curious to watch it, know that you have been warned.
Aliens and bikini armor aside, it is amazing how Japan is always finding ways to keep its culture alive. Because of that, even prehistoric artifacts such as dogū still have a place in modern Japan – and not only a place in museums, as national treasures, but also as pop culture icons.
Habu, J. (2004) Ancient Jomon of Japan. Cambridge University Press, Cambridge.
Henshall, K.G. (2004) A History of Japan: From Stone Age to Superpower. Second Edition. Palgrave Macmillan, Hampshire.
Imamura, K. (1996) Prehistoric Japan: New Perspectives on Insular East Asia. University of Hawaii Press, Honolulu.
Kaner, S. (2009) The Power of Dogu: Ceramic Figures from Ancient Japan. British Museum Press, London.
Kaner, S. (2011) The archaeology of religion and ritual in the prehistoric Japanese archipelago. In: Insoll, T. (Ed.) The Oxford Handbook of the Archaeology of Ritual and Religion. Oxford University Press, Oxford. Pp. 457–469.
Those figures presented here that were extracted from the Tokyo National Museum (Digital Research Archives: http://webarchives.tnm.jp/) and Wikimedia Commons, have been slightly modified (cropped, etc.) to improve presentation.
ABOUT THE AUTHOR
Dr. Rodrigo Salvador is a paleontologist and biologist, but is irredeemably fascinated with archaeology and mythology. Although his main “thing” remains Ancient Egypt, he is becoming increasingly drawn to the Jōmon and Yayoi periods of Japanese history. He has faced Japanese pre-historic monsters in many JRPGs, sometimes even summoning them to fight on his behalf – well, actually that last bit was just in SMT/Persona, because who on Earth uses a Claydol?
 Back then, in my home country, Internet connection was awfully slow and the service very expensive.
 The phallic stone rods seen above (Fig. 5) are also typically regarded as fertility symbols (Habu, 2004).
 That happened to other weird beings, such as the cartoonish Egyptian god Medjed (Salvador, 2017).
 And talking about exaggerating dates, the Japanese archaeologist Shinichi Fujimura claimed to have found Paleolithic artifacts in Japan dating back to 600,000 years ago. However, it was later discovered that he fabricated his own artifacts and planted them on his excavation site so he could “find” them later (Romey, 2001; Normile, 2001).
 Arahabaki’s look was very different in early SMT games, such as Megami Tensei II, where it was depicted as a samurai of sorts. So maybe they just retained the name, alongside the original idea/description, and changed this monster’s appearance to that of a dogū in later games.
Like most regular children in the 2000’s, we were obsessed with Pokémon games and anime series. The experience of exploring new environments, discovering new creatures and collecting them, always fascinated us. Maybe this was a sign of what we would become: zoologists. During college, as we got to know ever more about animal biodiversity, we could not help but notice several similarities between Pokémon and real animals. Today, as an arachnologist and an entomologist, and still Pokémon fans, our interest in arthropods and admiration for this franchise were the main motivations for this study.
ANIMALS IN THE MEDIA
Animal diversity has always been debated and represented in different types of media. Since the Pleistocene, humans depict animals in their paintings (Aubert et al., 2014), likely reflecting an age-old fascination with nature that still endures. Or, as E. O. Wilson puts it in his “biophilia hypothesis”: “humans have an innate desire to catalog, understand, and spend time with other life-forms” (Wilson, 1984). Given this, studies relating Zoology and culture, especially pop culture, are becoming more and more common recently. Just to name some examples including arthropods, Coelho (2000, 2004) studied insect references in lyrics and cover art of rock music albums, Castanheira et al. (2015) analyzed the representation of arthropods in cinematographic productions, Salvador (2016) studied the biology of giant centipedes in the Gears of War game franchise, and Da-Silva & Campos (2017) analyzed the representation of ants in the Ant-Man movie. There are even some science outreach works about the Pokémon franchise as the analysis of the ichthyological diversity in the Pokémon world (Mendes et al., 2017) and the study of the group of birds popularly called “robins” represented in the game (Tomotani, 2014).
Arthropods correspond to the largest part of the known biotic diversity in the world, counting with over 80% of animal diversity (Zhang, 2011a). With lots of morphological variation, the phylum Arthropoda is divided into five subphyla: Trilobitomorpha (the trilobites, now extinct); Chelicerata (arachnids, horseshoe crabs, and others); Crustacea (shrimps, lobsters, crabs, barnacles and woodlice); Hexapoda (insects) and Myriapoda (centipedes and millipedes). With a high biomass, terrestrial arthropods can be easily seen in a variety of environments, and their presence affects us in several ways.
Although arthropods can inspire fear as venomous creatures or disease vectors, actually most of them are either harmless or important for our own well-being and survival. For instance, many groups of insects are extremely important pollinators and without them, agriculture would collapse. Moreover, terrestrial arthropods have a considerable role as bioindicators for assessing environmental quality (Andersen, 1990; Brown, 1997; Fischer, 2000; Ferrier et al., 2004) and some even have remarkable medicinal uses (Kumar et al., 2015).
POKÉMON, A BRIEF STORY
The word “Pokémon” is a contraction from the Japanese “Pocket Monsters” (ポケモン). The idea consists in fictional creatures – the eponymous Pokémon – that humans can capture and train to do all sorts of chores, the main one of which is fighting each other. Created by Satoshi Tajiri, Pokémon was originally a game released in 1996, but its tremendous success soon spawned an anime series, mangas, animated movies, a card game, and countless ”goodies” (toys, accessories, clothing, candies, etc.). Developed by Game Freak and published by Nintendo, today Pokémon is one of the most successful game franchises in history, with more than 270 million of overall game copies sold around the world (The Pokémon Company, 2017).
The anime series was released in 1997 and was an instant success with kids, remaining so to this day. Many episodes have an environmental tone, showing how humans can affect the habitats and biodiversity of Pokémon, and emphasizing the importance of collecting for species preservation (Bainbridge, 2013). As a game franchise, Pokémon reached mainly teenagers, which remains a loyal customer base to this day. Today, the games are in their seventh generation (“Gen VII”) and each generation adds a new territory to be explored and several new creatures to be caught. As of now, there are 802 creatures, but some new ones have already been announced for the second game of Gen VII.
The creator of Pokémon, Satoshi Tajiri, loved to collect bugs when he was young, which likely influenced his creation. The Pokémon are mostly inspired by animals and plants and some of them have particular features that can be related to certain real species. In this way, Pokémon biodiversity can be seen as a virtual sample of natural biodiversity.
The main objective of this study is to survey all Pokémon inspired by arthropods, up to Gen VII, and conduct a comparative biological classification of them until the taxonomic level of “Order”, if possible. Considering the Pokémon world as a simulation of our own natural world, we also investigate if the different arthropod groups have the same real-world representativeness in Pokémon. This can be done by analyzing the proportion of species of each group.
MATERIAL AND METHODS
The sources of information used for this study are: Bulbapedia (https://bulbapedia. bulbagarden.net) and The Official Pokémon Website (https://www.pokemon.com). The Pokémon were classified by Type, Generation, and by their respective taxonomic levels in real-world Biology: Phylum, Subphylum, Class and Order.
The classification into real-world taxonomic levels was made by analyzing morphological and behavioral characters present in the Pokémon species, and comparing them to the relevant animal groups (Fig. 1). Morphological characters were obtained by observing official illustrations and game models. Behavioral characters were obtained from the Pokédex entries of each Pokémon species. Some Pokémon species presented arthropod’s features that were too imprecise to be related to a certain subphyla or order, or their design included features from more than one group of arthropods (for instance, Venonat and Whirlipede). In these cases, the species were marked as “undetermined Subphylum/Order”; regardless, we always classified them to the most accurate level possible.
The biodiversity data used for comparison to the natural world were retrieved from Zhang (2011b).
We found a total of 91 Pokémon species inspired by arthropods, representing 11.3% of all Pokémon creatures. Most of them (19) belongs to Gen III, corresponding to 14.1% of the total in this generation (Fig. 2, Table 1).
Most of the Pokémon species could be classified into the four main living subphyla of Arthropoda: Hexapoda (Figs. 3A–H), Crustacea (Figs. 3I–M), Chelicerata (Figs. 3N–R) and Myriapoda (Figs. 3S–U). The three exceptions were: Kabutops, Anorith and Armaldo (Figs. 3V–X). The former was allocated to the entirely fossil subphylum Trilobitomorpha. The latter two were allocated into another fossil group, with an uncertain position inside Arthropoda (or even an external group, according to some researchers). They belong to the Class Dinocaridida, Order Radiodonta (this ranking is still highly debated, though) and are popularly known as “terror shrimps”.
The Arthropoda subphylum that inspired most of the Pokémon species was Hexapoda, with 62 pokémon, followed by Crustacea (12), Chelicerata (11) and Myriapoda (3) (Figs. 4–5).
The taxonomical order that inspired most of the arthropod Pokémon was Lepidoptera, represented by 21 species. This can be explained by the huge visual appeal and beauty of butterflies and moths. This explanation can be also applied to the large number of Pokémon inspired by the order Coleoptera (13 species), the beetles, animals with an astounding variation of colors and shape. The third order in diversity is Decapoda (10 species), represented by crabs and shrimps.
POKÉMON DIVERSITY vs NATURAL DIVERSITY
The large number of Pokémon inspired by Hexapoda is congruent with the high diversity of this group in the natural world (Table 2). The fact that there was more Pokémon inspired in Crustacea (Table 3) than in Chelicerata (Table 4) is at odds with natural diversity, but can be related to the very frequent contact that Japanese people have with aquatic animals, which are one of the country’s main food sources (Ashkenazi & Jacob, 2003). The few specimens of Myriapoda in the game are proportionally congruent with their diversity in nature (Table 5).
The comparison between natural and Pokémon diversity shows that the Pokémon world presents higher representativeness of arthropod-like creatures that are more familiar to people or that have a greater visual appeal. The latter is the case of Lepidoptera (Fig. 5), whose diversity in the Pokémon world is much higher than the second place (Coleoptera). However, beetles are the most diverse insect (and overall animal) group in the real world, with approximately 387,000 species, while lepidopterans count “just” with around 157,000 species (Zhang, 2011b). Proportionally, butterflies and moths represent 33.9% of Hexapoda in Pokémon, while in nature this percentage is much closer to that of Coleoptera within Hexapoda (37.6%) rather than the proportion of Lepidoptera (15.3%) (Table 2).
The large number of Pokémon inspired by arthropods indicates that this group, even though not as charismatic as mammalians or birds, still plays an important role in pop culture. The visual appeal and the everyday contact seems to be important aspects that ensure a higher diversity to certain arthropod-like groups in Pokémon. Nevertheless, the Pokémon world still seems to be a good virtual sample of the natural world and this kind of representation can be an interesting source for educational purposes, helping young people to know other type of animals that they do not usually have much contact with, including extinct species.
Aubert, M.; Brumm, A.; Ramli, M.; Sutikna, T.; Saptomo, W. E.; Hakim, B.; Morwood, J. M.; van den Bergh, D.G.; Kinsley, L.; Dossseto, A. (2014) Pleistocene cave art from Sulawesi, Indonesia. Nature 514: 223–227.
Andersen, A.N. (1990) The use of ant communities to evaluate change in Australian terrestrial ecosystems, a review and a recipe. Proceedings of the Ecological Society of Australia 16: 347–357.
Ashkenazi, M. & Jacob, J. (2003) Food Culture in Japan. Greenwood Press, Westport.
Bainbridge, J. (2013) “Gotta catch ‘em all!” Pokémon, cultural practice and object networks. IAFOR Journal of Asian Studies 1(1): 1–15.
Brown, K.S. (1997) Diversity, disturbance, and sustainable use of Neotropical forests: insects as indicators for conservation monitoring. Journal of Insect Conservation 1: 25–42.
Castanheira, P.S; Prado, A.W.; Da-Silva, E.R. (2015) Analyzing the 7th art – arthropods in movies and series. International Refereed Research Journal 3(1): 1–15.
Chatterton, B.D.E. (1994) Ordovician proetide trilobite Dimeropyge, with a new species from northwestern Canada. Journal of Paleontology 68(3): 541–556.
Coelho, J.R. (2000) Insects in Rock and Roll music. American Entomologist 46(3): 186–200.
Coelho, J.R. (2004) Insects in Rock and Roll cover art. American Entomologist 50(3): 142–151.
Collins, D. (1996) The “evolution” of Anomalocaris and its classification in the arthropod class Dinocarida (nov.) and order Radiodonta (nov.). Journal of Paleontology 70(2): 280–293.
Da-Silva, E.R. & Campos T.R.M. (2017) Ants in the Ant-Man movie, with biological notes. Journal of Geek Studies 4(2): 21–30.
Ferrier, S.; Powell, N.V.G.; Richardson, S.K.; Manion, G.; Overton, M.J.; Allnutt, F.T.; Cameron, E.S.; Mantle, K.; Burgess, D.N.; Faith, P.D. (2004) Mapping more of terrestrial biodiversity for global conservation assessment. Bioscience 54: 1101–1109.
Fischer, M. (2000) Species loss after habitat fragmentation. Trends in Ecology & Evolution 15: 396.
Kumar, V.; Roy, S; Sahoo, A.K.; Behera, B.K.; Sharma, A.P. (2015) Horseshoe crab and its medicinal values. International Journal of Current Microbiology and Applied Sciences 4 (1): 956–964.
Mendes, A.B.; Guimarães, F.V.; Eirado-Silva, C.B.P.; Silva, E.P. (2017) The ichthyological diversity of Pokémon. Journal of Geek Studies.4(1): 39–67.
Pokémon Company, The. (2017) Pokémon in Figures. Available from: http://www.pokemon. co.jp/corporate/en/data/ (Date of access: 15/ Sep/2017).
Salvador, R.B. (2016) The biology of giant war centipedes. Journal of Geek Studies 3(1): 1–11.
Wilson, E.O. (1984) Biophilia. Harvard University Press, Cambridge.
Zhang, Z.-Q. (2011a) Animal biodiversity: an introduction to higher-level classification and taxonomic richness. Zootaxa 3148: 7–12.
Zhang, Z.-Q. (2011b) Animal biodiversity: an outline of higher-level classification and survey of taxonomic richness. Zootaxa 3703: 1–82.
ABOUT THE AUTHORS
André Prado has a bachelor’s degree in Biological Sciences by UFRJ (Rio de Janeiro) and a master’s degree in Zoology by Museu Nacional (Rio de Janeiro). He is a great enthusiast of Cultural Zoology, studying especially the role of animals in cinema.
Thiago Avelar has a licentiate degree in Biological Sciences by UFRJ (Rio de Janeiro) and is currently a high school teacher (Colégio e Curso Miguel Couto, Rio de Janeiro). He was a Fairy Type Elite Four in the extinct Pokémon League Brazil.
 Biological classification organizes species into groups. From the largest to the smallest group: Domain, Kingdom, Phylum, Class, Order, Family, Genus, Species. Sometimes subcategories can exist inside one of these, like a “Subphylum” or “Subspecies”.
Pokémon, or Pocket Monsters, was originally created for videogames, becoming a worldwide fever among kids and teenagers in the end of the 1990’s and early 2000’s. Currently, it is still a success, with numerous games, a TV series, comic books, movies, a Trading Card Game, toys and collectibles. Through its core products and vibrant merchandising, Pokémon took over the world, influencing pop culture wherever it landed. Despite losing some steam in the early 2010’s, Pokémon is now back to its previous uproar with the release of Pokémon GO, an augmented reality (AR) game for smartphones. This game launched in 2016, with almost 21 million users downloading it in the very first week in the United States alone (Dorward et al., 2017). Thus, Pokémon is indubitably an icon in pop culture (Schlesinger, 1999a; Tobin, 2004).
The origin of Pokémon goes back to two role-playing video games (created by Satoshi Tajiri and released by Nintendo for the Game Boy; Kent, 2001): Pokémon Green and Pokémon Red, released in Japan in 1996. In the West, the Green version never saw the light of day, but the Red and Blue versions were released in 1998, selling together more than 10 million copies. Also in 1998, the Yellow version of the game was released, which has as its most distinct feature the possibility of having Pikachu (the most famous Pokémon) walking side by side with the player in the game. Pokémon Green, Red, Blue and Yellow are the so-called “first generation” of games in the franchise. Today, the Pokémon series is in its seventh generation, with 29 main games released, besides several spin-offs. The TV series, on the other hand, is in its sixth season, with more than 900 episodes.
The games and TV series take place in regions inhabited by many Pokémon and humans. The mission of the protagonist is to win competitions (“Pokémon battles”) against gym leaders who are spread across different cities and regions. For each victory, the protagonist receives a gym badge; with eight badges, he/she is allowed to enter the Pokémon League to try and become the Champion.
For each generation, new Pokémon (and an entire new region) are introduced. In this way, the creatures have a homeland, although most can appear in other regions as well (Schlesinger, 1999b; Whitehill et al., 2016). The seven main regions are: Kanto, Johto, Hoenn, Sinnoh, Unova, Kalos and Alola.
In every region, there are numbered routes that connect cities and landmarks and in which the protagonist travels, finding the monsters in their natural habitats and interacting with other characters. These routes comprise a great range of environments, such as forests, caves, deserts, mountains, fields, seas, beaches, underwater places, mangroves, rivers and marshes, which usually display a huge diversity of Pokémon.
In addition to winning the Pokémon League, the protagonist must complete the Pokédex, a digital encyclopedia of Pokémon. In other words, the trainer must catch all the Pokémon that live in that region, registering each capture in the Pokédex. Each Pokémon has a registry number and an entry text in the Pokédex. Pokémon are usually found in nature, and may be captured with a device called “Pokéball”. Pokéballs are small enough to fit in a pocket, hence the name “Pocket Monsters” (Whitehill et al., 2016).
NOT AS MONSTRUOUS AS WE THINK
In the world depicted in the games, there are 801 Pokémon, belonging to one or two of the following 18 types: Normal, Fire, Fighting, Water, Flying, Grass, Poison, Electric, Ground, Psychic, Rock, Ice, Bug, Dragon, Ghost, Dark, Steel and Fairy (Bulbapedia, 2017). Almost all Pokémon are based on animal species, some of them are based on plants or mythological creatures, and a few are based on objects. Curiously, all Pokémon are oviparous, which means they all lay eggs (their development happens inside of an egg and outside of their mother’s body); of course, in the real natural world, this is a reproductive strategy of animals such as fishes, amphibians, reptiles, birds and many kinds of invertebrates (Blackburn, 1999). Moreover, Pokémon might “evolve”, usually meaning they undergo some cosmetic changes, become larger and gain new powers.
In the present work, the Pokémon world was approached by analogies with the real natural world, establishing parallels with actual animals.
A remarkable group of animals represented in Pokémon is the fishes. Fishes are the largest group of vertebrates, with more than 32,000 species inhabiting marine and freshwater environments, a number that roughly corresponds to half of all described vertebrates (Nelson et al., 2016). Showing ample morphological and behavioral variety and living in most of the aquatic ecosystems of the planet, fishes are well represented in the Pokémon world, therefore offering a great opportunity for establishing parallels between the two worlds. The creators of the games not only used the morphology of real animals as a source of inspiration for the monsters, but also their ecology and behavior.
Based on these obvious connections between real fishes and Pokémon, the aim of this work is to describe the ichthyological diversity found in Pokémon based on taxonomic criteria of the classification of real fishes. Ultimately, our goal is to offer useful material for both teaching and the popularization of science.
Table 1. Taxonomic classification of the fish Pokémon. Abbreviations: Ch = Chondrichthyes; Gn = Gnathostomata; Pe = Petromyzontomorphi; Pt = Petromyzontida; Os = Osteichthyes. All images obtained from The Official Pokémon Website (2016).
GOTTA CATCH ‘EM FISHES!
The first step of our research was a search in the Pokédex (The Official Pokémon Website, 2016) for Pokémon which were related to fishes. The criterion used was the Pokémon’s morphology (resemblance to real fishes). Afterwards, the “fish Pokémon” were classified to the lowest taxonomic level (preferably species, but when not possible, genus, family or even order).
This classification of the Pokémon allowed the comparison of biological data (such as ecological, ethological, morphological traits) from Bulbapedia (2017) with the current knowledge on real fishes (e.g., Nelson et al., 2016). Bulbapedia is a digital community-driven encyclopedia created in 2004 and is the most complete source regarding the pocket monsters.
The final step was a search in online scientific databases (Fishbase, Froese & Pauly, 2016; and Catalog of Fishes, Eschmeyer et al., 2016) in order to obtain the current and precise taxonomy and additional information on habitats, ecology etc. of the fish species.
In the present work, the taxonomic classification used was that proposed by Nelson et al. (2016), who consider the superclasses Petromyzontomorphi (which includes the class Petromyzontida, that is, the lampreys) and Gnathostomata (the jawed vertebrates). Gnathostomata, in turn, includes the classes Chondrichthyes (cartilaginous fishes) and Osteichthyes (bony fishes). Along with this classification, we used the classification proposed by the database ITIS (Integrated Taxonomic Information System, 2016) for comparison at all taxonomic levels. Following identification, the “fish Pokémon” were described regarding their taxonomic and ecological diversity.
As a result of our search, 34 fish Pokémon were identified (circa 4% of the total 801 Pokémon; Table 1) and allocated in two superclasses, three classes, eighteen orders, twenty families and twenty-two genera. Eighteen of the 34 fish Pokémon (circa 53%) could be identified to the species level (Table 2). The features of the real fishes which probably inspired the creation of the Pokémon and other relevant information are described below for each species. To enrich the comparisons, images of the Pokémon (obtained from the Pokédex of The Official Pokémon Website; http://www.pokemon.com) and of the real fishes (illustrations by one of us, C.B.P. Eirado-Silva) follow the descriptions.
Table 2. Taxonomic diversity of the fish Pokémon.
Horsea and Seadra
Species:Hippocampus sp.; Common name: seahorse.
The Pokémon Horsea and Seadra (Fig. 1), which debuted in the first generation of the franchise, were based on seahorses. The long snout, ending in a toothless mouth (Foster & Vincent, 2004), the prehensile, curved tail (Rosa et al., 2006) and the salient abdomen are features of the real fishes present in these Pokémon. Seahorses belong to the genus Hippocampus, presently composed of 54 species (Nelson et al., 2016). The males have a pouch in their bellies where up to 1,000 eggs are deposited by the females. In this pouch, the eggs are fertilized and incubated for a period ranging from 9 to 45 days (Foster & Vincent, 2004). Due to overfishing for medicinal and ornamental purposes, as well habitat destruction, about 33 species of seahorses are considered threatened (Rosa et al., 2007, Castro et al., 2008; Kasapoglu & Duzgunes, 2014).
Figure 1. Horsea, Seadra and Hippocampus sp.
Goldeen and Seaking
Species:Carassius auratus; Common name: goldfish.
Goldeen and Seaking (Fig. 2) were based on the goldfish. This species is one of the most common ornamental fishes worldwide (Soares et al., 2000; Moreira et al., 2011) and it is widely used in studies of physiology and reproduction due to its docile behavior and easy acclimatization to artificial conditions (Bittencourt et al., 2012; Braga et al., 2016). The resemblance between the goldfish and the Pokémon include morphological features, such as the orange/reddish color and the long merged fins, and the name “Goldeen”. The name Seaking, on the other hand, may be a reference to another common name of the species, “kinguio”, from the Japanese “kin-yu” (Ortega-Salas & Reyes-Bustamante, 2006).
Figure 2. Goldeen, Seaking and Carassius auratus.
Species:Cyprinus carpio; Common name: common carp.
Possibly the most famous fish Pokémon, Magikarp (Fig. 3) was based on a common carp, a species present in Europe, Africa and Asia, widely used in pisciculture due to its extremely easy acclimatization to many freshwater environments and the high nutritional value of its meat (Stoyanova et al., 2015; Mahboob et al., 2016; Voigt et al., 2016). In some regions of the planet, such as Brazil, the common carp is considered an invasive species, as it was inadvertently released in the wild and poses a threat to the native aquatic fauna (Smith et al., 2013; Contreras-MacBeath et al., 2014).
Figure 3. Magikarp and Cyprinus carpio.
The shared traits between the Pokémon and the real fish are many: the rounded mouth, the lips, the strong orange color and the presence of barbels (“whiskers”) (Nelson et al., 2016). In China, the common carp is praised as an animal linked to honor and strength, due of its ability to swim against the current; an ancient legend tells about carps that swim upstream, entering through a portal and transforming into dragons (Roberts, 2004). In Pokémon, Magikarp evolves into Gyarados, which resembles a typical Chinese dragon.
Chinchou and Lanturn
Species:Himantolophus sp.; Common name: footballfish.
Chinchou and Lanturn (Fig. 4) were based on fishes of the genus Himantolophus, a group of deep-sea fishes found in almost all oceans living in depths up to 1,800 meters (Klepadlo et al., 2003; Kharin, 2006). These fishes are known as footballfishes, a reference to the shape of their bodies. Fishes of this genus have a special modification on their dorsal fin that displays bioluminescence (the ability to produce light through biological means; Pietsch, 2003), which is used to lure and capture prey (Quigley, 2014). Bioluminescence was the main inspiration for these Pokémon, which have luminous appendages and the Water and Electric types. The sexual dimorphism (difference between males and females) is extreme in these fishes: whilst females reach up to 47 cm of standard-length (that is, body length excluding the caudal fin), males do not even reach 4 cm (Jónsson & Pálsson, 1999; Arronte & Pietsch, 2007).
Figure 4. Chinchou, Lanturn and Himantolophus sp.
Species:Diodon sp.; Common name: porcupinefish.
Qwilfish (Fig. 5) was based on porcupinefishes, more likely those of the genus Diodon, which present coloring and spines most similar to this Pokémon. Besides the distinctive hard, sharp spines (Fujita et al., 1997), porcupinefishes have the ability to inflate as a strategy to drive off predators (Raymundo & Chiappa, 2000). As another form of defense, these fishes possess a powerful bacterial toxin in their skin and organs (Lucano-Ramírez et al., 2011; Ravi et al., 2016). Accordingly, Qwilfish has both Water and Poison types.
Figure 5. Qwilfish and Diodon sp.
Species:Remora sp.; Common names: remora, suckerfish.
Remoraid was based on a remora (Fig. 6), a fish with a suction disc on its head that allows its adhesion to other animals such as turtles, whales, dolphins, sharks and manta rays (Fertl & Landry, 1999; Silva & Sazima, 2003; Friedman et al., 2013; Nelson et al., 2016). This feature allows the establishment of a commensalisc or mutualisc relationship of transportation, feeding and protection between the adherent species and its “ride” (Williams et al., 2003; Sazima & Grossman, 2006). The similarities also include the name of the Pokémon and the ecological relationship they have with other fish Pokémon: in the same way remoras keep ecological relationships with rays, Remoraid does so with Mantyke and Mantine (Pokémon based on manta rays; see below).
Figure 6. Remoraid and Remora sp.
Mantyke and Mantine
Species:Manta birostris; Common name: manta ray.
The Pokémon Mantyke and its evolved form Mantine (Fig. 7) were probably based on manta rays of the species Manta birostris, which inhabits tropical oceans (Duffy & Abbot, 2003; Dewar et al., 2008) and can reach more than 6 meters of wingspan, being the largest species of ray in existence (Homma et al., 1999; Ari & Correia, 2008; Marshall et al., 2008; Luiz et al., 2009; Nelson et al., 2016). The similarities between the Pokémon and the real fish are: the body shape, the color pattern, the large and distinctive wingspan and even the names.
Figure 7. Mantine, Mantyke and Manta birostris.
Kingdra and Skrelp
Species:Phyllopteryx taeniolatus; Common name: common seadragon.
Kingdra and Skrelp (Fig. 8) were based on the common seadragon. The resemblances between these Pokémon and the real fish species include the leaf-shaped fins that help the animals to camouflage themselves in the kelp “forests” they inhabit (Sanchez-Camara et al., 2006; Rossteuscher et al., 2008; Sanchez-Camara et al., 2011), and the long snout. Also, the secondary type of Kingdra is Dragon. Although both are based on the common seadragon, Kingdra and Skrelp are not in the same “evolutionary line” in the game.
Common seadragons, as the seahorses mentioned above, are of a particular interest to conservationists, because many species are vulnerable due to overfishing, accidental capture and habitat destruction (Foster & Vincent, 2004; Martin-Smith & Vincent, 2006).
Figure 8. Kingdra, Skrelp and Phyllopteryx taeniolatus.
Species:Pygocentrus sp.; Common name: red piranha.
Piranhas of the genus Pygocentrus possibly were the inspiration for the creation of Carvanha (Fig. 9), a Pokémon of voracious and dangerous habits. The main feature shared by the real fish and the Pokémon is the color pattern: bluish in the dorsal and lateral areas, and reddish in the ventral area (Piorski et al., 2005; Luz et al., 2015).
It is worthwhile pointing out that, despite what is shown in movies and other media, piranhas do not immediately devour their prey; instead, they tear off small pieces, bit by bit, such as scales and fins (Trindade & Jucá-Chagas, 2008; Vital et al., 2011; Ferreira et al., 2014).
Figure 9. Carvanha and Pygocentrus sp.
Order: Carcharhiniformes; Common name: shark.
Sharpedo (Fig. 10), according to its morphological traits (elongated fins), was possibly based on sharks of the order Carcharhiniformes, the largest group of sharks, with 216 species in 8 families and 48 genera. Fishes in this order are common in all oceans, in both coastal and oceanic regions, and from the surface to great depths (Gomes et al., 2010). Several species of Carcharhiniformes are in the IUCN’s (International Union for Conservation of Nature) endangered species list (a.k.a. “Red List”) due to overfishing, as their fins possess high commercial value (Cunningham-Day, 2001).
Figure 10. Sharpedo and a carcharhiniform shark.
Species:Misgurnus sp.; Common name: pond loach.
Barboach (Fig. 11) is likely based on fishes of the genus Misgurnus, natively found in East Asia (Nobile et al., 2017) but introduced in several countries (Gomes et al., 2011). These animals, like M. anguillicaudatus Cantor, 1842, are used as ornamental fishes and in folk medicine (Woo Jun et al., 2010; Urquhart & Koetsier, 2014). The shared similarities between the Pokémon and the pond loach include morphological features, such as the elongated body, oval fins and the presence of barbels (Nelson et al., 2016). The resemblance also extends itself to behavior, such as the habit of burying in the mud (Zhou et al., 2009; Kitagawa et al., 2011) and using the barbels to feel the surroundings (Gao et al., 2014). The secondary type of Barboach, Ground, alongside the ability to feel vibrations in the substrate, seem to be a reference to the behavior of the real fishes.
Figure 11. Barboach and Misgurnus sp.
Species:Silurus sp.; Common name: catfish.
Whiscash (Fig. 12) was based on the Japanese mythological creature Namazu, a gigantic catfish that inhabits the underground realm and is capable of creating earthquakes (Ashkenazi, 2003). Namazu also names the Pokémon in the Japanese language (“Namazun”). In Japan, fishes of the genus Silurus are usually associated with this mythological creature and even the common name of these fishes in that country is “namazu” (Yuma et al., 1998; Malek et al., 2004). In addition, the physical traits of the Silurus catfishes also present in Whiscash are the long barbels (or “whiskers”, hence the name Whiscash) and the robust body (Kobayakawa, 1989; Kiyohara & Kitoh, 1994). In addition to the Water type, Whiscash is also Ground type, which is related to Namazu’s fantastic ability of creating earthquakes.
Figure 12. Whiscash and Silurus sp.
Species:Micropterus salmoides; Common name: largemouth bass.
The Pokémon Feebas (Fig. 13), a relatively weak fish (as its name implies), was possibly based on a largemouth bass, a freshwater fish native to North America (Hossain et al., 2013). The species was introduced in many countries and is often considered a threat to the native fauna (Welcomme, 1992; Hickley et al., 1994; Godinho et al., 1997; García-Berthou, 2002). Similarities between Feebas and the largemouth bass include the large, wide mouth and the brownish coloration, with darker areas (Brown et al., 2009).
Figure 13. Feebas and Micropterus salmoides.
Species:Regalecus sp.; Common name: oarfish.
Often praised as the most beautiful Pokémon of all (Bulbapedia, 2017), Milotic (Fig. 14) certainly lives up to its title. Their long reddish eyebrows were based on the first elongated rays of the dorsal fin of Regalecus species (Nelson et al., 2016), which also share the reddish color of the dorsal fin (Carrasco-Águila et al., 2014). Other similarities between the oarfish and the Pokémon are the elongated body (some oarfishes can grow larger than 3.5 m) and the spots scattered on the body (Chavez et al., 1985; Balart et al., 1999; Dulčić et al., 2009; Ruiz & Gosztonyi, 2010).
Figure 14. Milotic and Regalecus sp.
Species:Monognathus sp.; Common name: onejaw.
Probably based on fishes of the genus Monognathus, which have a large mandible and a long dorsal fin (Nelson et al., 2016), Huntail (Fig. 15) is one of the possible evolutionary results of the mollusk Pokémon Clamperl (the other possibility is Gorebyss; see below). According to Raju (1974), fishes of the genus Monognathus live in great depths and have a continuous dorsal fin that ends in an urostyle (“uro” comes from the Greek language and means “tail”, an element also present in the Pokémon’s name).
Figure 15. Huntail and Monognathus sp.
Family: Nemichthyidae; Common name: snipe eel.
The serpentine body and the thin beak-shaped jaw of Gorebyss (Fig. 16) are features of fishes belonging to the family Nemichthyidae (Nielsen & Smith, 1978). These fishes inhabit tropical and temperate oceans and can be found in depths up to 4,000 meters, in the so-called “abyssal zone” (Cruz-Mena & Anglo, 2016). The Pokémon’s name may be a reference to such habitat.
Figure 16. Gorebyss and a nemichthyid fish.
Species:Latimeria sp.; Common name: coelacanth.
Relicanth (Fig. 17) was based on the coelacanth. The brown coloration, the lighter patches on the body (Benno et al., 2006) and the presence of paired lobed fins (Zardoya & Meyer, 1997) are traits of both the real fish and the Pokémon. It was believed that coelacanths went extinct in the Late Cretaceous, but they were rediscovered in 1938 in the depths off the coast of South Africa (Nikaido et al., 2011). Therefore, the only two living species L. chalumnae Smith, 1939 and L. menadoensis Pouyaud et al., 1999 are known as “living fossils” (Zardoya & Meyer, 1997). Probably for this reason, Relicanth belongs to the Water and Rock types (the “fossil Pokémon” are all Rock-type).
Figure 17. Relicanth and Latimeria sp.
Species:Helostoma temminckii; Common name: kissing gourami.
The silver-pinkish coloration, the peculiar mouth formed by strong lips and the habit of “kissing” other individuals of their species (which is actually a form of aggression!) are features of the kissing gourami (Sterba 1983; Sousa & Severi 2000; Sulaiman & Daud, 2002; Ferry et al., 2012) that are also seen in Luvdisc (Fig. 18). Helostoma temminckii is native to Thailand, Indonesia, Java, Borneo, Sumatra and the Malay Peninsula (Axelrod et al., 1971), but due to its use an ornamental fish and the irresponsible handling by fishkeepers, it has been introduced in other parts of the world (Magalhães, 2007).
Figure 18. Luvdisc and Helostoma temminckii.
Finneon and Lumineon
Species:Pantodon buchholzi; Common name: freshwater butterflyfish.
Finneon and Lumineon (Fig. 19) were probably based on the freshwater butterflyfish. Finneon has a caudal fin in the shape of a butterfly and Lumineon, like Pantodon buchholzi, has large pectoral fins (Nelson et al., 2016) resembling the wings of a butterfly (hence the popular name of the species). Butterflyfishes are found in West African lakes (Greenwood & Thompson, 1960); their backs are olive-colored while their ventral side is silver, with black spots scattered throughout the body; their fins are pink with some purplish spots (Lévêque & Paugy, 1984). Both Pokémon have color patterns that resemble the freshwater butterflyfish.
Figure 19. Finneon, Lumineon and Pantodon buchholzi.
Family: Serrasalmidae; Common name: piranha.
The two forms of the Pokémon Basculin (Fig. 20) seem to have been inspired on fishes from the Serrasalmidae family, such as piranhas. Basculin, like these fishes, has a tall body and conical teeth (Baumgartner et al., 2012). Piranhas are predators with strong jaws that inhabit some South American rivers. Curiously, they are commonly caught by local subsistence fishing (Freeman et al., 2007).
Figure 20. Basculin’s two forms and a serrasalmid fish.
Species:Mola mola; Common name: sunfish.
The very name of this Pokémon is evidence that it was inspired on Mola mola, the sunfish (Fig. 21). Moreover, Alomomola, just like the sunfish, has a circular body with no caudal fin (Pope et al., 2010). The sunfish is the largest and heaviest bony fish in the world, weighting more than 1,500 kg (Freesman & Noakes, 2002; Sims et al., 2009). They inhabit the Atlantic and Pacific Oceans, feeding mainly on zooplankton (Cartamil & Lowe, 2004; Potter & Howell, 2010).
Figure 21. Alomomola and Mola mola.
Tynamo, Eelektrik and Eelektross
Species: Petromyzon marinus; Common name: sea lamprey.
The evolutionary line Tynamo, Eelektrik and Eelektross (Fig. 22) was probably inspired by the life cycle of the sea lamprey, Petromyzon marinus: Tynamo represents a larval stage, Eelektrik a juvenile, and Eelektross an adult. As a larva, the sea lamprey inhabits freshwater environments and, after going through metamorphosis, the juvenile migrates to the ocean, where they start to develop hematophagous (“blood-sucking”) feeding habits (Youson, 1980; Silva et al., 2013). Eelektrik and Eelektross, like the sea lamprey, have a serpentine body and a circular suction cup-mouth with conical teeth. In addition, the yellow circles on the side of these Pokémon resemble the gill slits of lampreys (which are of circular shape) or the marbled spots of P. marinus (Igoe et al., 2004).
It is worth mentioning that Eelektrik and Elektross also seem to possess name and characteristics (Electric type and serpentine body with yellow spots) inspired by the electric eel (Electrophorus electricus Linnaeus, 1766), a fish capable of generating an electrical potential up to 600 volts, making it the greatest producer of bioelectricity in the animal kingdom (Catania, 2014). However, a remarkable characteristic of Eelektrik and Eelektross is the jawless mouth structure of the superclass Petromyzontomorphi species. The electric eel has a jaw and thus belongs to the superclass Gnathostomata (jawed vertebrates) (Gotter et al., 1998).
Figure 22. Tynamo, Eelektrik, Eelektross and P. marinus.
Order: Pleuronectiformes; Common name: flatfish.
Flattened and predominantly brown in color, Stunfisk (Fig. 23) appears to have been based on fishes of the order Pleuronectiformes. Popularly known as flatfishes, these animals have both eyes on the same side of the head and stay most of their lives buried and camouflaged on sandy and muddy substrates of almost every ocean, feeding on fishes and benthic invertebrates (Sakamoto, 1984; Kramer, 1991; Gibb, 1997). It is likely that the primary type of Stunfisk, Ground, is based on the close relationship between pleuronectiform fishes and the substrate they live in. Species of this group are very valuable for the fishing industry (Cooper & Chapleau, 1998).
Figure 23. Stunfisk and a pleuronectiform fish.
Species:Phycodurus eques; Common name: leafy seadragon.
Dragalge (Fig. 24), a Pokémon belonging to the Poison and Dragon types, was based on a leafy seadragon. This species is found in Australia and it is named after its appearance: this fish has appendages throughout its body that resemble leaves (Larson et al., 2014). This feature, also present in the Pokémon, allows the leafy seadragon to camouflage itself among algae (Wilson & Rouse, 2010). Dragalge is the evolved form of Skrelp, a Pokémon based on a common seadragon (see above).
Figure 24. Dragalge and Phycodurus eques.
Species:Sardinops sagax; Common name: Pacific sardine.
Wishiwashi (Fig. 25) was probably based on the Pacific sardine, a pelagic fish with high commercial value and quite abundant along the California and Humboldt Currents (Coleman, 1984; Gutierrez-Estrada et al., 2009; Demer et al., 2012; Zwolinski et al., 2012). The lateral circles of the Pokémon are a reference to the dark spots present on the lateral areas of the real fish (Paul et al., 2001). Furthermore, Wishiwashi has the ability to form a large school, just as sardines do (Emmett et al., 2005; Zwolinski et al., 2007).
Figure 25. Wishiwashi and Sardinops sagax.
Another parallel is the geographic location: the Pokémon belongs to Alola, a fictional region based on Hawaii, and S. sagax is one of the most common sardines in the Eastern Pacific Ocean. From the mid-1920’s to the mid-1940’s, for example, S. sagax supported one of the largest fisheries in the world. The stock collapsed in the late 1940’s, but in the 1990’s it started to recover (McFarlane et al., 2005).
Species:Rhinecanthus rectangulus; Common name: reef triggerfish.
Bruxish (Fig. 26) was probably inspired by the species Rhinecanthus rectangulus, the reef triggerfish of the Hawaiian reefs and other tropical regions (Kuiter & Debelius, 2006; Dornburg et al., 2008). Bruxish has powerful jaws, just like the reef triggerfishes that prey upon a wide variety of invertebrates, such as hard-shelled gastropods, bivalves, echinoderms and crustaceans (Wainwright & Friel, 2000; Froese & Pauly, 2016).
Figure 26. Bruxish and Rhinecanthus rectangulus.
Besides the strong jaw, the overall body shape and the flashy coloring, another parallel can be seen: this Pokémon is an inhabitant of the Alola region (the Pokémon version of Hawaii) and R. rectangulus is actually the state symbol fish of the Hawaiian archipelago (Kelly & Kelly, 1997).
POCKET FISHES UNDER SCRUTINY
The majority of the identified Pokémon (85.29%) is, expectedly, Water-type. A large portion of them (29.41%) was introduced for the first time in the third generation of the franchise, in the Hoenn region.
Figure 27. Representativeness of fish classes in Pokémon.
Only three fish Pokémon were classified in the superclass Petromyzontomorphi (8.82%): the lamprey-like Tynamo, Eelektrik and Eelektross, all of them belonging to the same evolutionary line. In the superclass Gnathostomata, the class Osteichthyes is represented by the highest number of Pokémon: 28 in total (82.35%, Fig. 27). Inside this class, the most representative groups were the order Syngnathiformes (14.71%, Fig. 28), family Syngnathidae (15.63%, Fig. 29) and the genus Petromyzon (10.00%, Fig. 30).
Figure 28. Representativeness of fish orders in Pokémon.
Most of the real fishes on which the Pokémon were based (55.88%, Fig. 31) live in marine environments, followed by freshwater (continental water environments, 32.35%) and finally, brackish water (estuarine environments, 11.76%).
The “fish” species found in the Pokémon world consists of a considerable portion of the ichthyological diversity in our world. According to Nelson et al. (2016), the Osteichthyes class corresponds to 96.1% of all vertebrate fish species (30,508 species), followed by the Condrichthyes with 3.76% (1,197 species) and the Petromyzontida with just 0.14% (46 species). In Pokémon, the proportions of taxa (taxonomic group) that inspired the creatures follow a roughly similar distribution: within the 26 taxa in which the evolutionary families of the Pokémon were based, 23 are Osteichthyes class (88.46%), two are Condrichthyes (7.7%) and one is Petromyzontida (3.84%). If the games follow a pattern of introducing more fish Pokémon over time, it is expected that these proportions will gradually become more equivalent as each new generation of the franchise is released.
Figure 29. Representativeness of fish families in Pokémon.
ALMOST A BIOLOGICAL POCKET-WORLD
Our analysis shows that fish Pokémon are very diverse creatures, both taxonomic and ecologically, despite being a small group within the Pokémon universe (with 801 “species”).
The fish Pokémon are represented by several orders, families and genera of real fishes and, as previously stated, this is actually a relevant sampling of the ichthyological diversity of our planet. The marine Pokémon described here are inhabit from abyssal zones to coastal regions, including reefs. The creative process of the fish monsters in the game must have included a fair share of research on real animals.
Figure 30. Representativeness of fish genera in Pokémon.
The Hoenn region, which has the largest playable surface and includes areas with “too much water”, is also the region with the highest number of fish Pokémon. Furthermore, the majority of these Pokémon live in the marine environment and belongs to the Osteichthyes class, as is observed for real fishes (Nelson et al., 2016; Eschmeyer et al., 2016). However, it is also important to underline that marine fishes are those with the more attractive colors and shapes and, therefore, higher popular appeal, which is vital for a game based in charismatic monsters (Darwall et al., 2011; McClenachan, 2012; Dulvy et al., 2014).
Figure 31. Environments inhabited by the fish Pokémon.
In the present work, the analogy between fish Pokémon and real species allowed a descriptive study of the “Pokéfauna” in a similar manner in which actual faunal surveys are presented. These surveys are an important tool for understanding the structure of communities and to evaluate the conservation status of natural environments (Buckup et al., 2014). It is noteworthy that the association of the monsters with real fishes was only possible because Pokémon have several morphological, ecological and ethological traits that were based on real species.
Pokémon is a successful franchise and many of its staple monsters are already part of the popular imaginary. The creation of the pocket monsters was not done in a random manner; they were mostly inspired by real organisms, particularly animals, and often have specific biological traits taken from their source of inspiration. Thus, analogies between Pokémon and our natural world, such as the ones performed here, open a range of possibilities for science outreach.
Ari, C. & Correia, J.P. (2008) Role of sensory cues on food searching behavior of a captive Manta birostris (Chondrichtyes, Mobulidae). Zoo Biology 27(4): 294–304.
Arronte, J.C. & Pietsch, T.W. (2007) First record of Himantolophus mauli (Lophiiformes: Himantolophidae) on the slope off Asturias, Central Cantabrian Sea, Eastern North Atlantic Ocean. Cybium 31(1): 85–86.
Ashkenazi, M. (2003) Handbook of Japanese Mythology. ABC-CLIO, Santa Barbara.
Axelrod, H.R.; Emmens, C.W.; Sculthorpe, D.; Einkler, W.V.; Pronek, N. (1971) Exotic Tropical Fishes. TFH Publications, New Jersey.
Balart, E.F.; Castro-Aguirre, J.L.; Amador-Silva, E. (1999) A new record of the oarfish Regalecus kinoi (Lampridiformes: Regalecidae) in the Gulf of California, Mexico. Oceánides 14(2): 137–140.
Baumgartner, G.; Pavanelli, C.S.; Baumgartner, D.; Bifi, A.G.; Debona, T.; Frana, V.A. (2012) Peixes do Baixo Rio Iguaçu: Characiformes. Eduem, Maringá.
Benno, B.; Verheij, E.; Stapley, J.; Rumisha, C.; Ngatunga, B.; Abdallah, A.; Kalombo, H. (2006) Coelacanth (Latimeria chalumnae Smith, 1939) discoveries and conservation in Tanzania. South African Journal of Science 102: 486–490.
Bittencourt, F.; Souza, B.E.; Boscolo, W.E.; Rorato, R.R.; Feiden, A.; Neu, D.H. (2012) Benzocaína e eugenol como anestésicos para o quinguio (Carassius auratus). Arquivo Brasileiro de Medicina Veterinária e Zootecnia 64(6): 1597–1602.
Blackburn, D.G. (1999) Viviparity and oviparity: evolution and reproductive strategies. In: Knobil, E. & Neil, J. D. (Eds.) Encyclopedia of reproduction. Acedemic Press, New York. Pp. 994–1003.
Braga, W.F.; Araújo, J.G.; Martins, G.P.; Oliveira, S.L.; Guimarães, I.G. (2016) Dietary total phosphorus supplementation in goldfish diets. Latin American Journal of Aquatic Research 44(1): 129–136.
Brown, T.G.; Runciman, B.; Pollard, S.; Grant, A.D.A. (2009) Biological synopsis of largemouth bass (Micropterus salmoides). Canadian Manuscript Report of Fisheries and Aquatic Sciences 2884: 1–35.
Buckup, P.A.; Britto, M.R.; Souza-Lima, R.S.; Pascoli, J.C.; Villa-Verde, L.; Ferraro, G.A.; Salgado, F.L.K; Gomes, J.R. (2014) Guia de Identificação das Espécies de Peixes da Bacia do Rio das Pedras, Município de Rio Claro, RJ. The Nature Conservancy, Rio de Janeiro.
Carrasco-Águila, M.A.; Miranda-Carrillo, O.; Salas-Maldonado, M. (2014) El rey de los arenques Regalecus russelii, segundo ejemplar registrado en Manzanillo, Colima. Ciencia Pesquera 22(2): 85–88.
Cartamil, D.P. & Lowe, C.G. (2004) Diel movement patterns of ocean sunfish Mola mola off southern California. Marine Ecology Progress Series 266: 245–253.
Castro, A.L.C.; Diniz, A.F.; Martins, I.Z.; Vendel, A.L.; Oliveira, T.P.R.; Rosa, I.M.L. (2008) Assessing diet composition of seahorses in the wild using a nondestructive method: Hippocampus reidi (Teleostei: Syngnathidae) as a study-case. Neotropical Ichthyology 6(4): 637–644.
Catania, K. (2014) The shocking predatory strike of the electric eel. Science 346(6214): 1231–1234.
Chávez, H.; Magaña, F.G.; Torres-Villegas, J.R. (1985) Primer registro de Regalecus russelii (Shaw) (Pisces: Regalecidae) de aguas mexicanas. Investigaciones Marinas CICIMAR 2(2): 105–112.
Coleman, N. (1984) Molluscs from the diets of commercially exploited fish off the coast of Victoria, Australia. Journal of the Malacological Society of Australia 6: 143–154.
Contreras-Macbeath, T.; Gaspar-Dillanes, M.T.; Huidobro-Campos, L.; Mejía-Mojica, H. (2014) Peces invasores em el centro de México. In: Mendoza, R. & Koleff, P. (Eds.) Especies Acuáticas Invasoras en México. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad, Ciudad de México. Pp. 413–424.
Cooper, J.A. & Chapleau, F. (1998) Monophyly and intrarelationships of the family Pleuronectidae (Pleuronectiformes), with a revised classification. Fishery Bulletin 96(4): 686–726.
Cruz-Mena, O.I. & Angulo, A. (2016) New records of snipe eels (Anguilliformes: Nemichthyidae) from the Pacific coast of lower Central America. Marine Biodiversity Records 9(1): 1–6.
Cunningham-Day, R. (2001) Sharks in Danger: Global Shark Conservation Status with Reference to Management Plans and Legislation. Universal Plubishers, Parkland.
Darwall, W.R.T.; Holland, R.A.; Smith, K.G.; Allen, D.; Brooks, E.G.E.; Katarya, V.; Pollock, C.M.; Shi, Y.; Clausnitzer, V.; Cumberlidge, N.; Cuttelod, A.; Dijkstra, B.K.; Diop, M.D.; García, N.; Seddon, M.B.; Skelton, P.H.; Snoeks, J.; Tweddle, D.; Vié, J. (2011) Implications of bias in conservation research and investment for freshwater species. Conservation Letters 4: 474–482.
Demer, D.A.; Zwolinski, J.P.; Byers, K.A.; Cutter, G.R.; Renfree, J.S.; Sessions, T.S.; Macewicz, B.J. (2012) Prediction and confirmation of seasonal migration of Pacific sardine (Sardinops sagax) in the California Current Ecosystem. Fishery Bulletin 110(1): 52–70.
Dewar, H.; Mous, P.; Domeler, M.; Muljadi, A.; Pet, J.; Whitty, J. (2008) Movements and site fidelity of the giant manta ray, Manta birostris, in the Komodo Marine Park, Indonesia. Marine Biology 155(2): 121–133.
Dornburg, L.; Santini, F.; Alfaro, M.E. (2008) The influence of model averaging on clade posteriors: an example using the triggerfishes (family Balistidae). Systematic Biology 57(6): 905–919.
Dorward, L.J.; Mittermeier, J.C.; Sandbrook, C.; Spooner, F. (2017) Pokémon GO: benefits, costs, and lessons for the conservation movement. Conservation Letters 10(1): 160–165.
Duffy, C.A.J. & Abbott, D. (2003) Sightings of mobulid rays from northern New Zealand, with confirmation of the occurrence of Manta birostris in New Zealand waters. New Zealand Journal of Marine and Freshwater Research 37(4): 715–721.
Dulčić, J.; Dragičević, B.; Tutman, P. (2009) Record of Regalecus glesne (Regalecidae) from the eastern Adriatic Sea. Cybium 33(4): 339–340.
Emmett, R.L.; Blodeur, R.D.; Miller, T.W.; Pool, S.S.; Krutzikowsky, G.K.; Bentley, P.J.; McCrae, J. (2005) Pacific sardine (Sardinops sagax) abundance, distribution, and ecological relationships in the Pacific Northwest. California Cooperative Oceanic Fisheries Investigations Reports 46: 122–143.
Eschmeyer, W.N.; Fricke, R.; van der Laan, R. (2016) Catalog of Fishes: Genera, Species, References. Available from: http://researcharch ive.calacademy.org/research/ichthyology/catalog/fishcatmain.asp (Date of access: 25/Nov/ 2016).
Ferreira, F.S.; Vicentin, W.; Costa, F.E.S.; Suárez, Y.R. (2014) Trophic ecology of two piranha species, Pygocentrus nattereri and Serrasalmus marginatus (Characiformes, Characidae), in the floodplain of the Negro River, Pantanal. Acta Limnologica Brasiliensia 26(4): 381–391.
Ferry, L.A.; Konow, N.; Gibb, A.C. (2012) Are kissing gourami specialized for substrate-feeding? Prey capture kinematics of Helostoma temminckii and other anabantoid fishes. Journal of Experimental Zoology 9999A: 1–9.
Fertl, D. & Landry, A.M. Jr. (1999) Sharksucker (Echeneis naucrates) on a bottlenose dolphin (Tursiops truncatus) and a review of other cetacean-remora associations. Marine Mammal Science 15(3): 859–863.
Forsgren, K.L. & Lowe, C.G. (2006) The life history of weedy seadragons, Phyllopteryx taeniolatus (Teleostei: Syngnathidae). Marine and Freshwater Research 57: 313–322.
Foster, S.J. & Vincent, A.C.J. (2004) Life history and ecology of seahorses: implications for conservation and management. Journal of Fish Biology 65(1): 1–61.
Freedman, J.A. & Noakes, D.L.G. (2002) Why are there no really big bony fishes? A point-of-view on maximum body size in teleosts and elasmobranchs. Reviews in Fish Biology and Fisheries 12: 403–416.
Freeman, B.; Nico, L.G.; Osentoski, M.; Jelks, H.L.; Collins, T.M. (2007) Molecular systematics of Serrasalmidae: deciphering the identities of piranha species and unraveling their evolutionary histories. Zootaxa 1484: 1–38.
Friedman, M.; Johanson, Z.; Harrington, R.C.; Near, T.J.; Graham, M.R. (2013) An early fossil remora (Echeneoidea) reveals the evolutionary assembly of the adhesion disc. Proceedings of the Royal Society B 280(1766): 1–8.
Froese, R. & Pauly, D. (2016) FishBase, v. 10/2016. Available from: http://fishbase.org (Date of access: 25/Jan/2017).
Fujita, T.; Hamaura, W.; Takemura, A.; Takano, K. (1997) Histological observations of annual reproductive cycle and tidal spawning rhythm in the female porcupine fish Diodon holocanthus. Fisheries Science 63(5): 715–720.
Gao, L.; Duan, M.; Cheng, F.; Xie, S. (2014) Ontogenetic development in the morphology and behavior of loach (Misgurnus anguillicaudatus) during early life stages. Chinese Journal of Oceanology and Limnology 32(5): 973–981.
García-Berthou, E. (2002) Ontogenetic diet shifts and interrupted piscivory in introduced largemouth bass (Micropterus salmoides). International Review of Hydrobiology 87(4): 353–363.
Gibb, A.C. (1997) Do flatfish feed like other fishes? A comparative study of percomorph prey-capture kinematics. The Journal of Experimental Biology 200: 2841–2859.
Godinho, F.N.; Ferreira, M.T.; Cortes, R.V. (1997) The environmental basis of diet variation in pumpkinseed sunfish, Lepomis gibbosus, and largemouth bass, Micropterus salmoides, along an Iberian river basin. Environmental Biology of Fishes 50(1): 105–115.
Gomes, C.I.D.A.; Peressin, A.; Cetra, M.; Barrela, W. (2011) First adult record of Misgurnus anguillicaudatus Cantor, 1842 from Ribeira de Iguape River Basin, Brazil. Acta Limnologica Brasiliensia 23(3): 229–232.
Gomes, U.L.; Signori, C.N.; Gadig, O.B.F.; Santos, H.R.S. (2010) Guia para Identificação de Tubarões e Raias do Rio de Janeiro. Technical Books Editora, Rio de Janeiro.
Gotter, A.L.; Kaetzel, M.A.; Dedman, J.R. (1998) Electrophorus electricus as a model system for the study of membrane excitability. Comparative Biochemistry and Physiology 119A(1): 225–241.
Greenwood, P.H. & Thompson, K.S. (1960) The pectoral anatomy of Pantodon buchholzi Peters (a freshwater flying fish) and the related Osteoglossidae. Journal of Zoology 135: 283–301.
Gutiérrez-Estrada, J.C.; Yáñez, E.; Pulido-Calvo, I.; Silva, C.; Plaza, F.; Bórquez, C. (2009) Pacific sardine (Sardinops sagax Jenyns, 1842) landings prediction: a neural network ecosystemic approach. Fisheries Research 100: 116–125.
Hickley, P.; North, R.; Muchiri, S.M.; Harper, D.M. (1994) The diet of largemouth bass, Micropterus salmoides, in Lake Naivasha, Kenya. Journal of Fish Biology 44(4): 607–619.
Homma, K.; Maruyama, T.; Itoh, T.; Ishihara, H.; Uchida, S. (1999) Biology of the manta ray, Manta birostris Walbaum, in the Indo-Pacific. In: Séret, B. & Sire, J.-Y. (Eds.) Proceedings of the 5th Indo-Pacific Fish Conference. Ichthyological Society of France, Noumea. Pp. 209–216.
Hossain, M.M.; Perhar, G.; Arhonditsis, G.B.; Matsuishi, T.; Goto, A.; Azuma, M. (2013) Examination of the effects of largemouth bass (Micropterus salmoides) and bluegill (Lepomis macrochirus) on the ecosystem attributes of lake Kawahara-oike, Nagasaki, Japan. Ecological Informatics 18: 149–161.
Igoe, F.; Quigley, D.T.G.; Marnell, F.; Meskell, E.; O’Connor, W.; Byrne, C. (2004) The sea lamprey Petromyzon marinus (L.), river lamprey Lampetra fluviatilis (L.) and brook lamprey Lampetra planeri (Bloch) in Ireland: general biology, ecology, distribution and status with recommendations for conservation. Proceedings of the Royal Irish Academy 104B (3): 43–56.
ITIS. (2016) Integrated Taxonomic Information System. Available from: http://itis.gov/ (Date of access: 25/Nov/2016).
Jónsson, G. & Pálsson, J. (1999) Fishes of the suborder Ceratioidei (Pisces: Lophiiformes) in Icelandic and adjacent waters. Rit Fiskideildar 16: 197–207.
Kasapoglu, N. & Duzgunes, E. (2014) Some population characteristics of long-snouted seahorse (Hippocampus guttulatus Cuvier, 1829) (Actinopterygii: Syngnathidae) in the Southeastern Black Sea. Acta Zoologica Bulgarica 66(1): 127–131.
Kelly, S. & Kelly, T. (1997) Fishes of Hawaii: Coloring Book. Bess Press, Honolulu.
Kent, S.L. (2001) The Ultimate History of Video Games. The Crown Publishing Group, New York.
Kharin, V.E. (2006). Himantolophus sagamius (Himantolophidae), a new fish species for fauna of Russia. Journal of Ichthyology 46(3): 274–275.
Kitagawa, T.; Fujii, Y.; Koizumi, N. (2011) Origin of the two major distinct mtDNA clades of the Japanese population of the oriental weather loach Misgurnus anguillicaudatus (Teleostei: Cobitidae). Folia Zoologica 60(4): 343–349.
Kiyohara, S. & Kitoh, J. (1994) Somatotopic representation of the medullary facial lobe of catfish Silurus asotus as revealed by transganglionic transport of HRP. Fisheries Science 60(4): 393–398.
Klepladlo, C.; Hastings, P.A.; Rosenblatt, R.H. (2003) Pacific footballfish, Himantolophus sagamius (Tanaka) (Teleostei: Himantolophi-dae), found in the surf-zone at Del Mar, San Diego County, California, with notes on its morphology. Bulletin South California Academy of Sciences 102(3): 99–106.
Kobayakawa, M. (1989) Systematic revision of the catfish genus Silurus, with description of a new species from Thailand and Burma. Japanese Journal of Ichthyology 36(2): 155–186.
Kramer, S.H. (1991) The shallow-water flatfishes of San Diego County. California Cooperative Oceanic Fisheries Investigations Reports 32: 128–142.
Kuiter, R.H. & Debelius, H. (2006) World Atlas of Marine Fishes. Hollywood Import and Export, Frankfurt.
Larson, S.; Ramsey, C.; Tinnemore, D.; Amemiya, C. (2014) Novel microsatellite loci variation and population genetics within leafy seadragons, Phycodurus eques. Diversity 6: 33–42.
Lévêque, C. & Paugy, D. (1984) Guide des Poissons d’Eau Douce: de la Zone du Programme de Lutte contre l’Onchocercose em Afrique de l’Ouest. ORSTOM, Paris.
Lucano-Ramírez, G.; Peña-Pérez, E.; Ruiz-Ramírez, S.; Rojo-Vázquez, J.; González-Sansón, G. (2011) Reproducción del pez erizo, Diodon holocanthus (Pisces: Diodontidae) en la plataforma continental del Pacífico Central Mexicano. Revista de Biologia Tropical 59 (1): 217–232.
Luiz, O.J. Jr.; Balboni, A.P.; Kodja, G.; Andrade, M.; Marum, H. (2009) Seasonal occurrences of Manta birostris (Chondrichthyes: Mobulidae) in southeastern Brazil. Ichthyological Research 56(1): 96–99.
Luz, L.A.; Reis, L.L.; Sampaio, I.; Barros, M.C.; Fraga, E. (2015) Genetic differentiation in the populations of red piranha, Pygocentrus nattereri Kner (1860) (Characiformes: Serrasalminae), from the river basins of northeastern Brazil. Brazilian Journal of Biology 75(4): 838–845.
Magalhães, A.L.B. (2007) Novos registros de peixes exóticos para o Estado de Minas Gerais, Brasil. Revista Brasileira de Zoologia 24(1): 250–252.
Mahboob, S.; Kausar, S.; Jabeen, F.; Sultana, S.; Sultana, T.; Al-Ghanin, K.A.; Hussain, B.; Al-Misned, F.; Ahmed, Z. (2016) Effect of heavy metals on liver, kidney, gills and muscles of Cyprinus carpio and Wallago attu inhabited in the Indus. Brazilian Archives of Biology and Technology 59(e16150275): 1–10.
Malek, M.A.; Nakahara, M.; Nakamura, R. (2004) Uptake, retention and organ/tissue distribution of 137Cs by Japanese catfish (Silurus asotus Linnaeus). Journal of Environmental Radioactivity 77(2): 191–204.
Marshall, A.D.; Pierce, S.J.; Bennett, M.B. (2008) Morphological measurements of manta rays (Manta birostris) with a description of a foetus from the east coast of Southern Africa. Zootaxa 1717: 24–30.
Martin-Smith, K.M. & Vincent, A.C.J. (2006) Exploitation and trade of Australian seahorses, pipehorses, sea dragons and pipefishes (family Syngnathidae). Oryx 40(2): 141–151.
McClenachan, L.; Cooper, A.B.; Carpenter, K.E.; Dulvy, N.K. (2012) Extinction risk and bottlenecks in the conservation of charismatic marine species. Conservation Letters 5: 73–80.
McFarlane, G.A.; MacDougall, L.; Schweigert, J.; Hrabok, C. (2005) Distribution and biology of Pacific sardines (Sardinops sagax) off British Columbia, Canada. California Cooperative Oceanic Fisheries Investigations 46: 144–160.
Moreira, R.L.; da Costa, J.M.; Teixeira, E.G.; Moreira, A.G.L.; De Moura, P.S.; Rocha, R.S.; Vieira, R. H.S.F. (2011) Performance of Carassius auratus with diferent food strategies in water recirculation system. Archivos de Zootecnia 60(232): 1203–1212.
Nelson, J.S.; Grande, T.C.; Wilson, M.V.H. (2016) Fishes of the World. Wiley, New Jersey.
Nielsen, J.G. & Smith, D.G. (1978) The eel family Nemichthyidae (Pisces, Anguilliformes). Dana Report 88: 1–71.
Nikaido, M.; Sasaki, T.; Emerson, J.J.; Aibara, M.; Mzighani, S.I.; Budeba, Y.L.; Ngatunga, B.P.; Iwata, M.; Abe, Y.; Li, W.H.; Okada, N. (2011) Genetically distinct coelacanth population off the northern Tanzanian coast. Proceedings of the National Academy of Sciences of the United States 108(44): 18009–18013.
Nobile, A.B.; Freitas-Souza, D.; Lima, F.P.; Bayona Perez, I.L.; Britto, S.G.C.; Ramos, I.P. (2017) Occurrence of Misgurnus anguillicaudatus (Cantor, 1842) (Cobitidae) in the Taquari River, upper Paraná Basin, Brazil. Journal of Applied Ichthyology (in press).
Official Pokémon Website, The. (2016) The Official Pokémon Website. Available from: http://poke mon.com/ (Date of access: 20/Nov/2016).
Ortega-Salas, A.A. & Reyes-Bustamante, H. (2006) Initial sexual maturity and fecundity of the goldfish Carassius auratus (Perciformes: Cyprynidae) under semi-controlled conditions. Revista de Biologia Tropical 54(4): 1113–1116.
Paul, L.J.; Taylor, P.R.; Parkinson, D.M. (2001) Pilchard (Sarditlops neopilchardus) biology and fisheries in New Zealand, and a review of pilchard (Sardinops, Sardina) biology, fisheries, and research in the main world fisheries. New Zealand Fisheries Assessment Report 37: 1–44.
Pietsch, T.W. (2003) Himantolophidae. Footballfishes (deepsea anglerfishes). In: Carpenter, K.E. (Ed.) FAO Species Identification Guide for Fishery Purposes. The Living Marine Resources of The Western Central Atlantic. Vol. 2: Bony Fishes Part 1 (Acipenseridae to Grammatidae). Food and Agriculture Organization of the United Nations, Rome. Pp. 1060–1061.
Piorski, N.M.; Alves, J.L.R.; Machado, M.R.B.; Correia, M.M.F. (2005) Alimentação e ecomorfologia de duas espécies de piranhas (Characiformes: Characidae) do lago de Viana, estado do Maranhão, Brasil. Acta Amazonica 35(1): 63–70.
Pope, E.C.; Hays, G.C.; Thys, T.M.; Doyle, T.K.; Sims, D.S.; Queiroz, N.; Hobson, V.J.; Kubicek, L.; Houghton, J.D.R. (2010) The biology and ecology of the ocean sunfish Mola mola: a review of current knowledge and future research perspectives. Reviews in Fish Biology and Fisheries 20(4): 471–487.
Potter, I.F. & Howell, W.H. (2010) Vertical movement and behavior of the ocean sunfish, Mola mola, in the northwest Atlantic. Journal of Experimental Marine Biology and Ecology 396(2): 138–146.
Raju, S.N. (1974) Three new species of the genus Monognathus and the Leptocephali of the order Saccopharyngiformes. Fishery Bulletin 72(2): 547–562.
Ravi, L.; Manu, A.; Chocalingum, R.; Menta, V.; Kumar, V.; Khanna, G. (2016) Genotoxicity of tetrodotoxin extracted from different organs of Diodon hystrix puffer fish from South East Indian Coast. Research Journal of Toxins 8(1): 8–14.
Raymundo, A.R. & Chiappa, X. (2000) Hábitos alimentarios de Diodon histrix y Diodon holocanthus (Pisces: Diodontidae) en las costas de Jalisco y Colima, México. Boletín del Centro de Investigaciones Biológicas 34(2): 181–210.
Roberts, J. (2004) Chinese Mythology A to Z. Facts on File, New York.
Rosa, I.L.; Oliveira, T.P.R; Castro, A.L.C.; Moraes, L.E.S.; Xavier, J.H.A.; Nottingham, M.C.; Dias, T.L.P.; Bruto-Costa, L.V.; Araújo, M.E.; Birolo, A.B; Mai, A.C.G; Monteiro-Neto, C. (2007) Population characteristics, space use and habitat associations of the seahorse Hippocampus reidi (Teleostei: Syngnathidae). Neotropical Icthyology 5(3): 405–414.
Rosa, I.L.; Sampaio, C.L.S.; Barros, A.T. (2006) Collaborative monitoring of the ornamental trade of seahorses and pipefishes (Teleostei: Syngnathidae) in Brazil: Bahia state as a case study. Neotropical Icthyology 4(2): 247–252.
Rossteucher, S.; Wenker, C.; Jermann, T.; Wahli, T.; Oldenberg, E.; Schmidt-Posthaus, H. (2008) Severe scuticociliate (Philasterides dicentrarchi) infection in a population of sea dragons (Phycodurus eques and Phyllopteryx taeniolatus). Veterinary Pathology 45(4): 546–550.
Ruiz, A.E. & Gosztonyi, A.E. (2010) Records of regalecid fishes in Argentine Waters. Zootaxa 2509: 62–66.
Sakamoto, K. (1984) Interrelationships of the family Pleuronectidae (Pisces: Pleuronectiformes). Memoirs of Faculty of Fisheries of Hokkaido University 31(1/2): 95–215.
Sanchez-Camara, J. & Booth, D.J. (2004) Movement, home range and site fidelity of the weedy seadragon Phyllopteryx taeniolatus (Teleostei: Syngnathidae). Environmental Biology of Fishes 70(1): 31–41.
Sanchez-Camara, J.; Booth, D.J.; Murdoch, J.; Watts, D.; Turon, X. (2006) Density, habitat use and behaviour of the weedy seadragon Phyllopteryx taeniolatus (Teleostei: Syngnathidae) around Sydney, New South Wales, Australia. Marine and Freshwater Research 57: 737–745.
Sanchez-Camara, J.; Booth, D.J.; Turon, X. (2005) Reproductive cycle and growth of Phyllopteryxtaeniolatus. Journal of Fish Biology 67(1): 133–148.
Sanchez-Camara, J.; Martin-Smith, K.; Booth, D.J.; Fritschi, J.; Turon, X. (2011) Demographics and vulnerability of a unique Australian fish, the weedy seadragon Phyllopteryx taeniolatus. Marine Ecology Progress Series 422: 253–264.
Sazima, I. & Grossman, A. (2006) Turtle riders: remoras on marine turtles in Southwest Atlantic. Neotropical Ichthyology 4(1): 123–126.
Schlesinger, H. (1999a) Pokémon Fever: The Unauthorized Guide. St. Martin’s Paperbacks, New York.
Schlesinger, H. (1999b) How to Become a Pokémon Master. St. Martin’s Paperbacks, New York.
Silva, S.; Servia, M.J.; Vieira-Lanero, R.; Barca, S.; Cobo, F. (2013) Life cycle of the sea lamprey Petromyzon marinus: duration of and growth in the marine life stage. Aquatic Biology 18: 59–62.
Silva-Jr., J.M. & Sazima, I. (2003) Whalesuckers and a spinner dolphin bonded for weeks: does host fidelity pay off? Biota Neotropica 3(2): 1–5.
Sims, D.W.; Queiroz, N.; Doyle, T.K.; Houghton, J.D.R.; Hays, G.C. (2009) Satellite tracking of the world’s largest bony fish, the ocean sunfish (Mola mola L.) in the North East Atlantic. Journal of Experimental Marine Biology and Ecology 370: 127–133.
Smith, W.S.; Biagioni, R.C.; Halcsik, L. (2013) Fish fauna of Floresta Nacional de Ipanema, São Paulo State, Brazil. Biota Neotropica 13(2): 175–181.
Soares, C.M.; Hayashi, C.; Gonçalves, G.S.; Galdioli, E.M.; Boscolo, W.R. (2000) Plâncton, Artemia sp., dieta artificial e suas combinações no desenvolvimento e sobrevivência do quinguio (Carassius auratus) durante a larvicultura. Acta Scientiarum 22(2): 383–388.
Sousa, W.T.Z. & Severi, W. (2000) Desenvolvimento larval inicial de Helostoma temminckii Cuvier & Valenciennes (Helostomatidae, Perciformes). Revista Brasileira de Zoologia 17(3): 637–644.
Sterba, G. (1983) The Aquarium Encyclopedia. MIT Press, Cambridge.
Stoyanova, S.; Yancheva, V.S.; Velcheva, I.; Uchikova, E.; Georgieva, E. (2015) Histological alterations in common carp (Cyprinus carpio Linnaeus, 1758) gills as potential biomarkers for fungicide contamination. Brazilian Archives of Biology and Technology 58(5): 757–764.
Sulaiman, Z.H. & Daud, H.K.H. (2002) Pond aquaculture of kissing gouramis Helostoma temminckii (Pisces: Helostomatidae) in Bukit Udal, Tutong: a preliminary investigation. Bruneiana 3: 34–41.
Tobin, J. (2004) Pikachu’s Global Adventure: The Rise and Fall of Pokémon. Duke University Press, Durham.
Trindade, M.E.J. & Jucá-Chagas, R. (2008) Diet of two serrasalmin species, Pygocentrus piraya and Serrasalmus brandtii (Teleostei: Characidae), along a stretch of the Rio de Contas, Bahia, Brazil. Neotropical Ichthyology 6(4): 645–650.
Urquhart, A.N. & Koetsier, P. (2014) Diet of a cryptic but widespread invader, the oriental weatherfish (Misgurnus anguillicaudatus) in Idaho, USA. Western North American Naturalist 74(1): 92–98.
Vital, J.F.; Varella, A.M.B.; Porto, D.B.; Malta, J.C.O. (2011) Sazonalidade da fauna de metazoários de Pygocentrus nattereri (Kner, 1858) no Lago Piranha (Amazonas, Brasil) e a avaliação de seu potencial como indicadora da saúde do ambiente. Biota Neotropica 11(1): 199–204.
Voigt, C.L.; Silva, C.P.; Campos, S.X. (2016) Avaliação da bioacumulação de metais em Cyprinus carpio pela interação com sedimento e água de reservatório. Química Nova 39(2): 180–188.
Wainwright, P.C. & Friel, J.P. (2000) Effects of prey type on motor pattern variance in tetraodontiform fishes. Journal of Experimental Zoology 286(6): 563–571.
Welcomme, R.L. (1992) A history of international introductions of inland aquatic species. ICES Marine Science Symposia 194: 3–14.
Whitehill, S.; Neves, L.; Fang, K.; Silvestri, C. (2016) Pokémon: Visual Companion. The Pokémon Company International / Dorling Kindersley, London.
Williams, E.H.; Mignucci-Giannoni, A.A.; Bunkley-Williams, L.; Bonde, R.K.; Self-Sullivan, C.; Preen, A.; Cockcroft, V.G. (2003) Echeneid-sirenian associations, with information on sharksucker diet. Journal of Fish Biology 63(5): 1176–1183.
Wilson, N.G. & Rouse, G.W. (2010) Convergent camouflage and the non-monophyly of ‘seadragons’ (Syngnathidae: Teleostei): suggestions for a revised taxonomy of syngnathids. Zoologica Scripta 39(6): 551–558.
Woo Jun, J.; Hyung Kim, J. Gomez, D.K.; Choresca, C.H. Jr.; Eun Han, J.; Phil Shin, S.; Chang Park, C. (2010) Occurrence of tetracycline-resistant Aeromonashydrophila infection in Korean cyprinid loach (Misgurnus anguillicaudatus). African Journal of Microbiology Research 4(9): 849–855.
Yuma, M.; Hosoya, K.; Nagata, Y. (1998) Distribution of the freshwater fishes of Japan: an historical review. Environmental Biology of Fishes 52(1): 97–124.
Zardoya, R. & Meyer, A. (1997) The complete DNA sequence of the mitochondrial genome of a “living fossil,” the coelacanth (Latimeria chalumnae). Genetics 146: 995–1010.
Zhou, X.; Li, M.; Abbas, K.; Wang, W. (2009) Comparison of haematology and serum biochemistry of cultured and wild dojo loach Misgurnus anguillicaudatus. Fish Physiology and Biochemistry 35(3): 435–441.
Zwolinski J.P.; Demer, D.A.; Byers, K.A.; Cutter, G.R.; Renfree, J.S.; Sessions, T.S.; Macewicz, B.J. (2012) Distributions and abundances of Pacific sardine (Sardinops sagax) and other pelagic fishes in the California Current Ecosystem during spring 2006, 2008, and 2010, estimated from acoustic-trawl surveys. Fishery Bulletin 110(1): 110–122.
Zwolinski, J.P.; Morais, A.; Marques, V.; Stratoudakis, Y.; Fernandes, P.G. (2007) Diel variation in the vertical distribution and schooling behaviour of sardine (Sardina pilchardus) off Portugal. Journal of Marine Science 64(5): 963–972.
Balmford, A.; Clegg, L.; Coulson, T.; Taylor, J. (2002) Why conservationists should heed Pokémon. Science 295: 2367.
Shelomi, M.; Richards, A.; Li, I.; Okido, Y. (2012) A phylogeny and evolutionary history of the Pokémon. Annals of Improbable Research 18(4): 15–17.
ABOUT THE AUTHORS
Augusto Mendes began his journey as a Pokémon trainer in his childhood, when his parents gave him a green Game Boy Color with Pokémon Red for Christmas. Currently, he is a master’s degree student in the Program of Marine Biology and Coastal Environments of UFF, where he works with zooarchaeology of fishes and education.
Felipe Guimarães is in love with Pokémon (since he first watched the TV series) and the natural world. He graduated in Biology from the UERJ, where he worked with taxonomy and ecology of fishes. He also works with popularization of science and environmental education.
Clara Eirado-Silva, when she was eight years old, told her parents she would study sharks. She has always been passionate about art too and draw since her childhood. Currently, she holds a “Junior Science” scholarship, working on fishing ecology with emphasis on reproductive biology. In her free time, she draws her much loved fishes.
Although Pokémon is not exactly Dr. Edson Silva’s cup of tea, he watched all movies with his daughter, who’s crazy about the little monsters. As fate would have it, his work on population genetics of marine organisms attracted a master’s student (A.B.M.) who’s an equally crazy pokéfan. May Arceus not spare him from the monsters!
Pokémon has been an enormous success around the globe for more than 20 years. In this paper, I tackle the “Who’s that Pokémon?” challenge from a machine learning perspective. I propose a machine learning pre-processing and classification pipeline, using convolutional neural networks for classification of Pokémon sprites.
COMPUTING AND IMAGE RECOGNITION
Since they were invented, computers became increasingly present in our everyday life. Initially restricted to mathematical problem-solving and military applications in ballistics and cryptography, their applications become more diverse by the minute. As of today, machines beat humans in lots of tasks, one of the most recent being AlphaGo’s victory over the Go world champion (Go Game Guru, 2017).
This achievement is a testament to the remarkable advances sustained by machines towards intelligent applications. Go, with its almost infinite combinations, is not an easy problem to solve by “brute force”, the strategy usually employed by computers against humans in other perfect information games.
But do not despair, for not all is lost in our fight against our future robot overlords, as computers still struggle with a task that humans were quite literally born to do: image and pattern recognition. However good a computer may be today, humans are still way better at noticing that, even though Figure 1 shows a car, something quite unusual happened to it.
Figure 1. Crashed car against a tree. This text was definitely not written by a robot overlord (yet). (Image extracted from Wikimedia Commons; Thue, 2005).
But computers are catching on! Advances in machine learning techniques, especially in supervised learning methods, and the ever-growing data available for feeding these algorithms have been enabling giant leaps in this field. In 2015, a 150 layers’ residual neural network ensemble, trained by the MSRA team, achieved a 62% average precision in the 2015 image classification challenge with a data set with more than 1,000 different objects (Large Scale Visual Recognition Challenge, 2015).
Figure 2. Some simple things may be hard to a computer. (“Tasks”; XKCD, available from https://xkcd.com/1425).
So, we wonder… How would our machines fare against a challenge tackled by children around the world for the last 22 years?
Figure 3. Who’s that Pokémon? (Screenshot from the Pokémon animated series.)
Pokémon is an extremely successful franchise of games and animated series targeted at young audiences (although some people, as the author, disagree with this classification). The franchise was created by Satoshi Tajiri in 1995, with the publishing of two games for Nintendo’s handheld console Game Boy. In the game, the player assumes the role of a Pokémon trainer, capturing and battling the titular creatures. It was an enormous success, quickly becoming a worldwide phenomenon (Wikipedia, 2017b).
The franchise started with a total of 151 monsters (Fig. 4), but today the games have reached their seventh iteration, counting with a total of 802 monsters.
Figure 4. Left to right: Bulbasaur, Charmander and Squirtle. (Official art by Ken Sugimori; image taken from Bulbapedia, 2017).
Each Pokémon belongs to one or two types indicating its “elemental affinity”, as well as its strengths and weakness against other types. This feature is essential to the gameplay, establishing a deep and complex rock-paper-scissor mechanic that lays at the foundation of the combat system. There are 18 types (they were only 15 in the first game), as seen in Figure 5 (Bulbapedia, 2017).
Figure 5. The 18 Pokémon types, depicted with their usual background colors.
In this paper, I examine the performance of convolutional neural networks (also known as ConvNets) in a Pokémon Type classification task given a Pokémon game sprite. I will present the data collected, the pre-processing and training pipelines, ending with the performance metrics of the selected model. All the data, implementation code and results, as well as a Jupyter Notebook with the explanation of all the steps, are available in a GitHub repository (https://github.com/hemagso/Neuralmon).
To train the models, I am going to use game sprites. The dataset (the sprite packs) was obtained at Veekun (2017). These packs contain sprites ripped from the games’ so-called generations 1 to 5. Although there have been new games (and new monsters) released since then, they use tridimensional animated models; making it harder to extract the resources from the games, as well as making it available in a format that can be fed to a machine learning method. As such, in this paper we will only use Pokémon up until the fifth generation of the games (649 in total).
Figure 6 depicts the sprites of the three first-generation starters throughout all the games considered in this study.
We can immediately see that detail level varies between games, due to the different hardware and capabilities of the gaming consoles. The first generation, released for Nintendo’s Game Boy, has almost no hue variation in a single sprite, although there is some hue information in the dataset (for instance, Bulbasaur is green, Charmander is red and Squirtle is blue; Fig. 6). As we go on, through Game Boy Advance to Nintendo DS, we see that the level of detail skyrockets, not only in terms of hue, but also in shapes.
At a first glance, we can also identify some typical problems encountered in image classification tasks. The images have different sizes. Even though the Aspect Ratio in all images stays at a one-to-one ratio, we have images ranging from 40-pixel width in the first generation to 96-pixel width in the fifth one (pay attention to the scales on the border on each sprite in Figure 6).
Figure 6. Example of the variation of the sprites for three Pokémon, as seen throughout games and generations.
Also, not all sprites fill the same space in each image. Sprites from the oldest generations seem to fill, in relative terms, a bigger portion of their images. This also happens within the same generation, especially in newer games, relating, in general, to the differences in size of each Pokémon and its evolutions (Fig. 7).
Figure 7. Bulbasaur’s evolutionary line, as seen in the game’s 5th generation. As the Pokémon evolves and gets larger, its sprite fills up a larger portion of the image.
To solve this problem, let’s apply some computer vision techniques to identify the main object in the image, delimitate its bounding box and center our image on that box. The pipeline for that is:
Convert the image to grayscale.
Apply a Sobel Filter on the image, highlighting the edges of the sprite. The Sobel filter is a 3×3 convolutional kernel (more about these handy little fellows later, but see also Sckikit-Image, 2017) that seeks to approximate the gradient of an image. For a given image ‘A’, the Sobel operator is defined as:
Fill the holes in the image, obtaining the Pokémon’s silhouette.
Calculate the Convex Hull of the silhouette, that is, the smallest convex polygon that includes all pixels from the silhouette.
Define the square bounding box from the convex hull calculated before.
Select the content inside the bounding box, and resize it to 64 x 64 pixels.
Figure 8. Examples of all steps of the sprite centering pipeline.
After following the pipeline outlined above, we obtain new sprites that maximize the filling ratio of the sprite on the image. Those steps were taken using skimage, an image processing library for the Python programming language. Figure 8 shows the results of our pipeline for the sprites of the three 1st generation starters and Venusaur.
Our proposed pipeline is extremely effective at the task at hand. That is to be expected, as our images are very simple sprites, with a very clear white background.
Finally, let’s apply our method on all our monsters and images. Figure 9 shows the results for a bunch of Pokémon.
Figure 9. Centering results over various 5th gen Pokémon.
Now that we have all our Pokémon images to build our image dataset, we have to classify them in accordance with the variable that we want to predict. In this paper, we will try to classify each Pokémon according to its correct type using only its image. For example, in Figure 10 we try to use the image inside the bounding box to classify the Pokémon in one of the 18 types, trying to match its true type (shown below each Pokémon).
Figure 10. Example Pokémon and their respective types. Top row: Ivysaur (left) and Pidgey (right). Bottom row: Steelix (left) and Lord Helix (right), praise be unto him.
But there is a catch. A significant portion of the Pokémon, like all those from Figures 9 and 10, have a dual type. That is, its true type will be a combination of two different types from that list of 18 types. In Figure 10, for instance, Ivysaur is both a Grass type and Poison type, and has the strengths and weakness of both types.
To take this into account, we would have to make our target classifications over the combination of types. Even if we disregard type order (that is, consider that a [Fire Rock] type is the same class as a [Rock Fire] one), we would end up with 171 possible classes. (Actually, this number is a little bit smaller, 154, as not all combinations exist in the games.)
To make things worse, some combinations are rare (Fig. 11), with only one or two Pokémon, thus limiting the available samples to learn from.
Figure 11. Some unique type combinations. Top row: Magcargo (left) and Surskit (right). Bottom row: Spiritomb (left) and Heatran (right).
Due to the reasons outlined above, I opted to disregard type combinations in this paper. As such, we are only taking into account the primary type of a Pokémon. For instance, in Figure 10 we would have: Ivyssaur: Grass; Pidgey: Normal; Steelix: Steel; Lord Helix: Rock.
I used a convolutional Neural Network as a predictor on our dataset. Neural networks are one among many kinds of predictive models usually used in machine learning, consisting of an interconnected network of simple units, known as Neurons. Based on a loose analogy with the inner workings of biological systems, Neural Networks are capable of learning complex functions and patterns through the combination of those simple units (Wikipedia, 2017a).
In its simplest form, a Neuron is nothing more than a linear function of its inputs, followed by a non-linear activation function (Fig. 12). However, through the combination of several layers, neural networks are capable of modelling increasingly complex relationships between the independent and dependent variables at hand (Fig. 13).
Figure 12. The basic unit of a Neural Network.
Figure 13. A slightly more complex architecture for a neural network, with one hidden layer.
Neural networks are not exactly new, as research exists since 1940 (Wikipedia, 2017a). However, only with recent computational advances, as well as the development of the backpropagation algorithm for its training, that its use became more widespread.
OK, this is enough to get us through the Neural Network bit. But what the hell “convolutional” means? Let’s first talk a little about Kernels.
In image processing, a Kernel (also known as Convolution Matrix or Mask) is a small matrix used in tasks as blurring, sharpening, edge detection, among others. The effect is obtained through the calculation of the matrix convolution against the appropriate Kernel, producing a new image. We have already seen a Kernel used in this paper, in our pre-processing pipeline, where we applied a Sobel Kernel to detect the edges of a sprite.
Figure 14. Sobel Kernel effect on Venusaur’s sprite.
The convolution operation may be thought of as a sliding of the Kernel over our image. The values in the Kernel multiply the values below them in the image, element-wise, and the results are summed to produce a single value of the convolution over that window. (A much better explanation about the convolution operation can be found at http://setosa.io/ev/ image-kernels/.) In Figure 15, we apply a vertical Sobel filter to detect sharp variations in color intensity (ranging in our grayscale images from 120 to 255).
Figure 15. Convolution example. The red area highlighted in the image is being convoluted with a Vertical Edge detector, resulting in the red outlined value on the resulting matrix.
But what the heck! What do those Kernels have to do with neural networks? More than we imagine! A convolutional layer of a neural network is nothing more than a clever way to arrange the Neurons and its interconnections to achieve an architecture capable of identifying these filters through supervised learning. (Again, a way better explanation about the whole convolutional network-stuff may be found in http://cs231n.github.io/convolutional-network s/.) In our pre-processing pipeline, we used a specific Kernel because we already knew the one that would excel at the task at hand, but in a convolutional network, we let the training algorithm find those filters and combine them in subsequent layers to achieve increasingly complex features.
Our Neural Network’s Architecture
I used a small-depth convolutional network for our Pokémon classification task (Fig. 16).
Figure 16. Architecture of the Neural Network used here.
Each layer of the image represents a layer in our convolutional network. After each layer, we obtain a state tensor that represents the output of that layer (the dimension of the tensor is listed on the right side of each layer).
A convolution layer then applies the convolution operation. In the first layer, we apply 32 kernels of size 5 to the input image, producing 32 outputs of size 60 x 60 (with each convolution the image size diminishes due to border effects).
We also use max polling layers that simply reduce a tensor region to a single one by getting its maximum value (Fig. 17). As such, after the application of a 2 x 2 max polling layer, we get a tensor that is a quarter of the size of the original.
Figure 17. Example of the max pooling operation.
At the end, we flatten our tensor to one dimension, and connect it to densely connected layers for prediction. Our final layer has size 18, the same size as the output domain.
Train and Validation
To achieve our model training we are going to split our dataset in two parts: (1) the ‘training dataset’ will be used by our training algorithm to learn the model parameters from the data; (2) the ‘validation dataset’ will be used to evaluate the model performance on unseen data. In this way, we will be able to identify overfitting issues (trust me, we are about to see a lot of overfitting).
But we can’t simply select a random sample of our sprites. Sprites from the same Pokémon in different games are very similar to each other, especially between games of the same generation (Fig. 18).
Figure 18. Sprites of Bird Jesus from Pokémon Platinum (left) and Diamond (right). Wait… was it the other way around?
Box 1. Performance Metrics
In this article, we used three performance metrics to assess our model performance:
(1) Accuracy: the percentage of predictions that got the right type classification of the Pokémon;
(2) Precision: the percentage of images classified as a class (type) that truly belonged to that class;
(3) Recall: the percentage of images of a class (type) that were classified as that class.
While accuracy enable us to get an overall quality of our model, precision and recall are used to gauge our model’s prediction of each class.
If we randomly select sprites, we incur on the risk of tainting our validation set with sprites identical to the ones on the training set, which would lead to a great overestimation of model performance on unknown data. As such, I opted for Pokémon-wise sample. That is, I assigned the whole Pokémon to a set, instead of assigning individual sprites. That way, if Charizard is assigned to the validation set, all its sprites would follow, eliminating the risk of taint.
I used 20% of the Pokémon for the test sample, and 80% for the training set, which leaves us with 2,727 sprites for training.
First Model: Bare Bones Training
For the first try, I fed the training algorithm the original sprites, while keeping the training/ validation split. The algorithm trained over 20 epochs, which took about a minute in total. The results obtained in this first training session are presented in Figure 19 (see also Box 1 for an explanation of the performance metrics).
Figure 19. Performance of the training set in the first try.
Impressive! We got all the classifications right! But are those metrics a good estimation of the model performance over unseen data? Or are those metrics showing us that our models learned the training sample by heart, and will perform poorly on new data? Spoiler alert: it will. Let’s get a good look at it: Figure 20 exhibits those same metrics for our validation set.
It seems that our model is indeed overfitting the training set, even if it’s performing better than a random guess.
Figure 20. Performance of the validation set in the first try.
But wait a minute… why haven’t we got any Flying type Pokémon? It turns out that there is only one monster with Flying as its primary type (Tornadus; Fig. 21), and he is included in the training set.
Figure 21. Tornadus is forever alone in the Flying type.
Second Model: Image Augmentation
The poor performance our first model obtained for the validation set is not a surprise. Image classification, as said in the introduction, is a hard problem for computers to tackle. Our dataset is too small and does not have enough variation to enable our algorithm to learn features capable of generalization over a wider application.
To solve at least part of the problem, let’s apply some image augmentation techniques. This involves applying random transformations over the training images, thus enhancing their variation. A human being would be able to identify a Pikachu, no matter its orientation (upside down, tilted to the side etc.) and we would like our model to achieve the same. As such, I applied the following range of transformations over our training dataset (Fig. 22): (1) random rotation up to 40 degrees; (2) random horizontal shifts up to 20% image width; (3) random vertical shifts up to 20% image height; (3) random zooming up to 20%; (4) reflection over the vertical axis; and (5) shear transformation over a 0.2 radians range.
Figure 22. Images obtained through the image augmentation pipeline for one of Bulbasaur’s sprites.
I applied this pipeline to all sprites in our training set, generating 10 new images for each sprite. This way, our training set was expanded to 27,270 images. But will it be enough? After training over 30 epochs (this time it took slightly longer, a little over 10 minutes in total), I obtained the following results (Fig. 23).
Figure 23. Performance of the training set for the second model.
Wait a minute, has our model’s performance decreased? Shouldn’t this image augmentation thing make my model better? Probably, but let’s not start making assumptions based on our training set performance. The drop in overall performance is due to the increase in variation in our training set and this could be good news if it translates into a better performance for the validation set (Fig. 24).
Figure 24. Performance of the validation set for the second model.
And here we have it! Image augmentation actually helped in the model’s performance. The accuracy was raised by 14 percentage points, to a total of 39%. We could keep trying to get a better model, fiddling with model hyper-parameters or trying net architectures, but we are going to stop here.
Taking a Closer Look on the Classifications
There are some things that I would like to draw your attention to. The types with greater prediction Accuracy are: Fire (61%), Water and Poison (54% each), Grass (47%), Electric (46%). The types with greater Recall (see Box 1) are: Dark (92%), Fire (74%), Water (55%), Normal (49%), Grass (42%).
It’s no surprise that the three main types (Fire, Water and Grass) are among the top five in both metrics. These types have very strong affinities with colors, an information easily obtained from the images. They also are abundant types, having lots of training examples for the model to learn from.
Now let’s look at some correctly and incorrectly classified Pokémon (Figs. 25 and 26, respectively).
Figure 26. Some incorrectly classified Pokémon. Top row: Mochoke (left), Our Good Lord Helix (center), Lugia (right). Bottom row: Gardevoir (left), Seviper (center), Vaporeon (right).
Even in this small sample, we can see that color plays an important part in the overall classification. For example, in the incorrectly-classified Pokémon, Machoke had good chances of being a Poison type, possibly due to its purple color. Likewise, Seviper was classified as a Dark type probably due to its dark coloration.
And why is that? Well, we may never know! One of the downsides of using deep neural networks for classification is that the model is kind of a “black box”. There is a lot of research going on trying to make sense of what exactly is the network searching for in the image. (I recommend that you search the Internet for “Deep Dream” for some very trippy images.)
For now, we can look at the first layer activations for some of the Pokémon and try to figure out what is it that each kernel is looking for. But as we go deeper into the network, this challenge gets harder and harder (Fig. 27).
Figure 27. First layer activations (partial) for the three 1st Gen starters.
39% accuracy may not seem that impressive. But an 18-class classification problem with as little data as this is a hard one, and our model achieves a 20 percentage points gain against a Zero Rule Baseline, which is to guess the most frequent class for all Pokémon. Table 1 lists the frequencies of each class on the test set, which gives us a 19.5% accuracy for Zero Rule.
Table 1. Type frequency for the test dataset.
But of course, we shouldn’t be measuring our machines against such clumsy methods if we expect them to one day become the dominant rulers of our planet, and computers still have a long way to go if they expect to beat my little brother in the “Pokémon Classification Challenge” someday. On the bright side, they probably already beat my old man. But this is a topic for another article…
Bulbapedia. (2017) Type. Available from: http:// bulbapedia.bulbagarden.net/wiki/Type (Date of access: 20/01/2017).
Go Game Guru. (2017) DeepMind AlphaGo vs Lee Sedol. Available from: https://gogameguru. com/tag/deepmind-alphago-lee-sedol/ (Date of access: 07/Mar/2017).
Henrique wants to be the very best, like no one ever was. When he isn’t playing games, devouring sci-fi literature or writing awesome articles to an obscure geek journal on the Internet, he works a full-time job applying machine learning to the banking industry. Sadly, he got misclassified by his own creation. – Grass? Come on!?
Gotta Train ’em All
I wanna be the very best / Like no one ever was
To model them is my real test / To train them is my cause
I will travel across the data / Searching far and wide
Each model to understand / The power that’s inside
Neural Net, gotta train ’em all / It’s you and me / I know it’s my destiny
Neural Net, oh, you’re my best friend / The world we must understand
Neural Net, gotta train ’em all / A target so true / Our data will pull us through
You teach me and I’ll train you
Neural Net, gotta train ’em all / Gotta train ’em all
 The exact date for the invention of the computer is quite difficult to pin down. Helpful devices for calculations have existed for centuries, but truly programmable computers are a recent invention. If we take as a cutoff criterion that the first computer must be Turing Complete (that is, being able to compute every Turing computable function), our first examples would be placed around the first half of the twentieth century. The first project of a Turing complete machine is attributed to Charles Babbage in the nineteenth century. His Analytical Engine, if ever built, would be a mechanical monstrosity of steel and steam that, although not very practical, would certainly be awesome.
 It is estimated that the game space of Go comprises around 2.08·10^170 legal positions or 208,168,199,381,979,984,699,478,633,344,862,770,286,522,453,884,530,548,425,(…)639,456,820,927,419,612,738,015,378,525,648,451,698,519,643,907,259,916,015,628,(…)128,546,089,888,314,427,129,715,319,317,557,736,620,397,247,064,840,935, if you want to be precise (Tromp & Farnebäck, 2016).
 Brute force search is a problem-solving strategy that consists in enumerating all possible solutions and checking which solves the problem. For example, one may try to solve the problem of choosing the next move in a tic-tac-toe game by calculating all possible outcomes, then choosing the move that maximizes the chance of winning.
 Ideally, we would split our dataset in 3 separate datasets: (1) the ‘training dataset’ would be used to learn the model coefficients; (2) the ‘validation dataset’ would be used to calibrate model hyperparameters, as the learning rate of the training algorithm or even the architecture of the model, selecting the champion model; (3) the ‘test dataset’ would be used to evaluate the performance of the champion model. That way, we avoid introducing bias in our performance estimates due to our model selection process. As we already have a way too small dataset (and we aren’t tweaking the model that much), we can disregard the test dataset.
 In machine learning context, an epoch corresponds to an iteration in which all the training data is exposed to the learning algorithm (not necessarily at once). In this case, the neural network learned from 20 successive iterations in which it saw all the data.
 I trained all models on Keras using the Tensorflow backend. The training was done in GPU, with a NVIDIA GTX 1080, on a PC running Ubuntu. For more details, see the companion Jupyter Notebook at GitHub (https://github. com/hemagso/Neuralmon).
In a previous paper, I presented a curious experiment of a fish playing Pokémon, made real and popular thanks to the wonders of the Internet (Tomotani, 2014). The Twitch Channel (Twitch, 2016), which sadly has been inactive for some time now, showed Grayson, the fish, playing Pokémon Red with the help of an image identification software, and was watched by millions of people (Johns, 2014). I showed that, when assuming that a fish player was the same as a random input of commands – a premise I do not find absurd – it would take quite a while to advance through a single route (although a very complex one) in the game (Fig. 1): circa 115,700 years.
Figure 1. Route 23; image taken from Tomotani (2014). The many ledges, which could make the path much more tortuous, made the Twitch Play community come up with the infamous name “Ledge Simulator”.
The peculiar premise of a fish playing Pokémon obviously derived from the original Twitch Plays Pokémon, a game of Pokémon Red where everyone watching the stream could type commands in the chat window. An IRC bot would read and execute the commands in the game. The available commands were the classic Gameboy keys: A, B, Up, Down, Left, Right, Start and Select.
Since the inputs came from rational human beings, with defined intentions and minimal coordination, supposedly the game should be less frustrating than watching a fish swimming and randomly inputting commands in the game – the key word in this phrase being “supposedly”.
The point is (besides, of course, the difficulty in coordinating thousands of people to avoid incorrect commands), not all of the people participating in the event wanted to complete the game. In fact, some of them wanted to make matters as difficult as possible for the other players, as their goal was solely to make the Twitch Plays Pokémon a frustrating experience for anyone wanting to be a crowdsourced Pokémon master. (In their “defense”, I find it hard to believe that this sort of behavior was not one of the intentions of the programmer of this game, described by him as a “social experiment”; Alcantara, 2014.) This group of people is given the name of trolls.
The impact of trolls on Twitch Plays Pokémon was so peculiar that it resulted in a curious work, where Machine Learning techniques were used to detect anomalous inputs in the game (from a base of 38 million data points), trying to identify potential trolls (Haque, 2014). The objective of the present study is to see exactly how the percentage of trolls inputting commands on Twitch Plays Pokémon affected the time for completing Route 23 in the game.
Trolls are creatures from Nordic and Scandinavian myths and folklore, made popular in the 20th century by pop culture, starting with J.R.R. Tolkien’s books, going through Dungeons & Dragons and Harry Potter to the thousands, possibly tens of thousands, of other novels, comics, games etc. which they inspired.
Troll is a term applied to the Giants of Norse Mythology (the Jötnar), a race that live in Jötunheimr, one of nine worlds in Norse cosmology. There is some confusion about the terms, though, as jötunn (the singular form of Jötnar), troll, þurs, and risi frequently overlap and are used to describe many beings in the legends. Some researchers point out that there are distinct classes of these creatures, but the terms are frequently considered synonyms in late medieval literature and all of them are frequently translated to English simply as “giant” (Jakobsson, 2005). In a late saga of the Icelanders (Bárðar saga Snæfellsáss; probably from the early 14th century), a passage at the very beginning claims that risi and troll not only are distinctive races, but are, respectively, at the opposite ends of the binary divide of good and evil (Jakobsson, 2005).
The Internet term “troll”, however, does not come from such creatures. The term “trolling” means luring others into pointless and time consuming discussions, deriving from the practice used in fishing where a baited line is dragged behind a boat (Herring et al., 2002).
Figure 2. Trollface, a popular internet meme based in “rage comics”. – Did you just read two paragraphs about creatures in Norse mythology that have absolutely nothing to do with this topic?
The idea of trolling always seems to be related to communication, mostly computer-mediated communication (CMC). Hardaker (2010) analyzed a 172-million-word corpus of unmoderated, asynchronous CMC to try to formulate an academic definition of trolling. After his analysis, he proposes that:
“A troller is a CMC user who constructs the identity of sincerely wishing to be part of the group in question, including professing, or conveying pseudo-sincere intentions, but whose real intention(s) is/are to cause disruption and/or to trigger or exacerbate conflict for the purposes of their own amusement. Just like malicious impoliteness, trolling can (1) be frustrated if users correctly interpret an intent to troll, but are not provoked into responding, (2) be thwarted, if users correctly interpret an intent to troll, but counter in such a way as to curtail or neutralize the success of the troller, (3) fail, if users do not correctly interpret an intent to troll and are not provoked by the troller, or, (4) succeed, if users are deceived into believing the troller’s pseudo-intention(s), and are provoked into responding. Finally, users can mock troll. That is, they may undertake what appears to be trolling with the aim of enhancing or increasing affect, or group cohesion.”
― Hardaker, 2010: p. 237.
The definition of “trolling” for the present study is not strictly the same, since the troll does not necessarily has the intention of portraying any good will toward the group’s goal of completing the Pokémon game. The intention of creating conflict and frustrating a group of people for personal amusement, though, is very similar. As such, we shall use a less strict definition of trolls, so that we may keep calling them such.
The percentage of people deliberately trolling on the Internet is never clear. Since trolls tend to draw too much attention, it is easy to believe they are more numerous. In the Twitch Plays Pokémon case, this was particularly true: a single input from a troll at the wrong time (or right time, from the troll’s point of view) and the avatar in the game would jump down a ledge, making many more inputs necessary for the avatar to go back to the same coordinate.
THE FIRST SIMULATION MODEL
For the present study, I wanted to know how trolls affected the time for completing Route 23 in Twitch Plays Pokémon. The first step was to develop a simulation model in VBA. A map for Route 23 composed of 305 different coordinates was generated (the same as seen in Tomotani, 2014) and the neighbors for each coordinate were defined. For each coordinate, I defined three different inputs: the “optimal route” (the command which a player wanting to finish the game would input) and two different “troll inputs”. The latter are commands that would make the route as long and frustrating as possible. I defined two different troll inputs because: (a) there were times when two commands could be equally bad; (b) trolls not always want to troll the same way; and (c) to add some variation to the routes’ heat maps presented in this analysis.
The premises to define the commands for the optimal route were:
Always go through the shortest path towards the objective (the door at the end of Route 23);
When two commands are equally good, stay away from ledges for as long as possible.
The premises to define the troll commands were as follows:
If you are close to a ledge you normally would not want to jump over, jump;
Go away from the direction you are supposed to go;
Only input “movement” commands (to simplify the model by not having to create simulation models for the game’s menu screen).
Once the simulation model was complete, I defined a “troll factor”, that is, the number of trolls inputting commands. Thus, the troll factor represents the percentage of players that are, in fact, trolls trying to prevent the group from completing the route.
In the simulation, we randomly decided whether the next command would be an “optimal” command or a “troll” command. The chance of the command being a “troll input” was equal to the “troll factor”. Figure 3 shows a heat map of time spent in each coordinate when the troll factor was zero. Since there is no probabilistic factor (that is, all commands made are optimal), this route is always the same: it takes 70 steps to complete this path.
Figure 3. Optimal route, when the troll factor is zero, completed in 70 steps.
Figure 4 shows a heat map of time spent in each coordinate for a simulation run when the troll factor was 10%, that is, on average one out of every ten inputs was made by a troll. In this specific run, it took 285 steps to complete the route, more than four times longer than the optimal path. When tested with a troll factor of 20%, the number of steps necessary reaches the thousands. Figure 5 is an example, where 1,011 steps had to be taken to complete the route.
Figure 4. Simulation run with troll factor of 10%, completed in 285 steps.
Figure 5. Simulation run with Troll factor of 20%, completed in 1,011 steps.
THE SECOND SIMULATION MODEL
The VBA model, though, proved inefficient when dealing with higher troll factors, constantly crashing or giving inconsistent results. As such, I decided to use a more appropriate tool, and developed a simple Discrete Event Simulation Model on the Rockwell Arena software (ARENA, 2016). This model can be seen in Figure 6. (For more information on simulations with this software, see Altiok & Melamed, 2007; and for more on Discrete Event Simulation, see Banks et al., 2009).
On this simulation model, the coordinates were indexed in a “305 lines x 4 columns” matrix in the software, where each line was a coordinate, and each column contained the possible neighbors. With a “Create” module, 100 entities were inserted simultaneously in the model, each representing a different simulation run. Each entity had an attribute named “position”, where the current coordinate of the simulation was recorded, and a “total steps” attribute, where the number of steps necessary to finish the simulation run was recorded.
Figure 6. Simulation model on Rockwell Arena.
In each step of the simulation, a “Decide” module of the Arena decided whether the next command inserted for each run was a “normal” or a “troll” one, and the “position” attribute of each entity was updated. When the current position of an entity was the coordinate for the door at the end of Route 23, the simulation was terminated and the total number of steps to finish the route was registered. At the end of the simulation, a “Read and Write” module was used to record some additional information at an Excel worksheet, such as number of steps on each cell, and number of steps on each area (more of this below).
To see how the percentage of trolls (“troll factor”) affected the number of steps, I made various simulations. For each troll factor used, I ran 100 simulations and calculated the average number of steps to complete Route 23 (it took a while for the 50% troll factor!). The results can be seen on Table 1 and Figure 7.
Table 1. Average number of steps (out of 100 simulations) necessary to finish Route 23.
Figure 7. Well, that escalated quickly. (Keep in mind that the Y axis is in logarithmic scale.)
In other words, with a single troll command in every 20, it is already enough to make traversing this map twice as difficult, and when 50% of the inputs were made by trolls (well… in this case it is almost a philosophical question whether the trolls are the ones trying to prevent others from completing the game or the ones effectively trying to complete it), the number of steps necessary was more than 264,000 times greater than the optimal route.
Additionally, I divided Route 23’s map into five different areas (Fig. 8) to analyze how much time was spent in each area for each troll factor. The results can be seen on Table 2 and Figure 9. It is clear that, the greater the troll factor, the easier it is for the avatar of the game to jump over the lower ledge and into “Area 5”, where he spends most of the time. (See the Appendix for heat maps with average results.)
Figure 8. Division of Route 23 into five areas.
Table 2. Percentage of time spent on each of the five Areas for varying troll factors.
Figure 9. Percentage of time spent in each of the five Areas for varying troll factors.
With these results, it is clear that trolls can be a pain whenever you are trying to conduct some nice experiment online, or have a good argument. Herring et al. (2002) speculate about why trolls (or “trollers”, as he puts it) troll, suggesting that the actions may be a result of: hatred towards people who are perceived as different or threatening by the troller (e.g., women or homosexuals); sense of control and self-empowerment when groups are targeted for their vulnerability (such as disabled people or inexperienced users); or simply because trollers enjoy the attention they receive, even (and maybe especially) when it is negative. According to Herring et al. (2002), this suggests that ignoring the troller might truly be an effective way of thwarting him/her (a.k.a. “don’t feed the troll”). Sadly, this is much harder to do in Twitch Plays Pokémon.
AND WHAT ABOUT THE FISH?
Now consider again my previous work (Tomotani, 2014), where I discussed the fish Grayson and his journey to be a Pokémon master. When you think about it, the average number of steps necessary for a simulation with 50% of trolls seems a bit underwhelming. Considering that one command was inputted every 1.5 second, the 18.492.842 steps would be made in 321 days, less than one year, while random commands made by a fish would take more than 115 thousand years.
I tried to validate this number by using my model to simulate Grayson, but it would take way too long. I adapted the model so that, instead of deciding between a “troll” and “normal” input, it chose randomly between any of the four directions (Fig. 10). After ninety minutes and 5 billion steps, the model seemed no closer to concluding its task. A simple calculation showed me it would take close to 30 days to simulate the equivalent of 115 thousand years (not that long in comparison to the 10 million years it took Earth to calculate the question to the ultimate answer).
Figure 10. Adapted model for the “random” input. Here, the two “troll inputs” and the “optimal input” from the previous model are substituted by four different commands, one for each direction.
So, I aborted the idea of simulating the whole thing, and decided to limit my simulation. I made 10 simultaneous runs, each with a limit of 1 billion steps, 50 times more than the average it took for the troll factor of 50%. After this limit was reached, the simulation would stop and record the results to show how far the simulation managed to go. Spoiler alert: not a single one managed to finish the route. Figure 11 shows a heat map for this experiment, considering the sum of the 10 simulations, a total of 10 billion steps. Figure 12 gives a closer look at the hardest area to traverse (the narrow path on Area 3), where a single input “down” means lots of backtracking.
After 10 billion steps, only once the random simulation managed to get to the “signpost” coordinate, and never getting further than that (you can see the actual signpost in Fig. 12). Table 3 shows the distribution of steps in each Area (of those five defined above) by the random simulation. More than 95% of the time was spent in Area 5, the lower part of the map.
Figure 11. Heat map for the “random” simulation.
Figure 12. Closer look at the most critical part of the map, a narrow path on Area 3. The number on each coordinate represents the number of steps taken on that coordinate (the “###” represents very large numbers).
Table 3. Steps taken in each area in “random” simulation.
As such, the conclusion of my previous work that a random input of commands to complete the game is not a productive approach seems valid. This might be a possible explanation to the deactivation of the Fish Plays Pokémon channel, so that Grayson could focus his energy in activities that make better use of his skills.
Banks, J.; Carson II, J.S.; Nelson, B.L.; Nicol, D.M. (2010) Discrete-Event System Simulation, 5th ed. Prentice Hall, Upper Saddle River.
Haque, A. (2014) Twitch Plays Pokemon, Machine Learns Twitch: unsupervised context-aware anomaly detection for identifying trolls in streaming data. Available from: https://www. albert.cm/dl/twitch_paper.pdf (Date of access: 19/Sep/2016).
Hardaker, C. (2010) Trolling in asynchronous computer-mediated communication: from user discussions to theoretical concepts. Journal of Politeness Research 6(2): 215–242.
Herring, S.; Job-Sluder, K.; Scheckler, R.; Barab, S. (2002) Searching for safety online: managing “trolling” in a feminist forum. Information Society 18(5): 371–384.
Jakobsson, A. (2005) The Good, the Bad, and the Ugly: Bárðar saga and its Giants. Medieval Scandinavia 15: 1–15.
Earlier this year, an article entitled “Which is The Most Offensively Powerful Starter Pokémon?” (Codd, 2016) caused great controversy on the Internet among players and fans of the Pokémon franchise. This article compared the three classical starter Pokémon, based on the anime, and concluded that Charizard was the strongest one.
The present work aims to analyze and discuss the data presented by Codd (2016) regarding the following issues: (1) Does his anime-based data coincide with the game mechanics? (2) Can his study be applied to metagame prospects? (3) Is Charizard really the most “powerful” Pokémon in-game?
Pokémon™ is an entertainment franchise, created by Satoshi Tajiri in 1995, that started with video games, but now includes an anime, a trading card game, clothing and several other products. Needless to say, the main products of the franchise (the games and anime) caused a large impact in recent pop culture.
The first products to be released were the “twin games” Pokémon Red and Pokémon Green, in 1996 in Japan. These games were later (in 1998) released worldwide as the Red and Blue versions for Nintendo’s Game Boy console. (As a side note, in celebration of its 20 years of existence, earlier this year the Pokémon Company released a website containing a timeline of their products).
On TV, Pokémon was first released in Japan in 1997 with the episode “Pokémon – I Choose You” (released in the United States only in 1998; Wikipedia, 2016), triggering wide public attention. The franchise is now successful worldwide, attracting millions of fans and players of all ages, ethnic groups and social classes, and the games are often regarded extremely seriously by the players.
Codd (2016) concluded in his article that Charizard (the last form of the starter Charmander) was the most powerful of the three initial options (the grass-type Bulbasaur, the fire-type Charmander and the water-type Squirtle; Fig. 1). To reach this conclusion, Codd based his work on “data” provided by the anime, specifically (for Charizard) in the episode “Can’t beat the heat!” (aired 17/Feb/2002), from which he estimated variables such as weight (body mass), height and width of the Pokémon. Through a series of calculations, all very well-founded in Physics, Codd determined that the offensive power of Charizard is well ahead of its competitors.
Codd’s calculations are in fact quite accurate and may be applicable to the anime. But it behooves us a little analysis regarding the applicability of his results to the game. At the very start of his article, Codd states:
“At the start of each Pokémon game, the player is given a choice of starter Pokémon. The options are almost always a choice between a fire type, a water type and a grass type. In most ways the most iconic of the starter Pokémon across all Pokémon generations are the original three; Charmander, Squirtle and Bulbasaur, which will fully evolve into Charizard, Blastoise and Venusaur respectively.”
― Codd (2016: p. 1), my highlight
Therefore, the first sentence of this quotation makes it clear that the author refers to the games, with its challenging proposition of having to choose one of three possible options to continue. In the same paragraph Codd says:
“Each of these Pokémon also have a signature move, one which is closely linked to them through the course of the anime and the games. For Charizard this is Flamethrower, for Blastoise this is Hydro Pump and for Venusaur this is Solar Beam.”
― Codd (2016: p. 1), my highlight
Thus, the author establishes an intrinsic connection between anime and game. From this point on, he starts his analysis based on the size and proportions of the starting Pokémon gathered from the anime. Despite this, the authors surmises that his calculations may be applied to the game. The discordance between Codd’s arguments and the games is based on a simple fact: he used estimates and variables that are not true (or accounted for) in the native mechanics of the game, being thus irrelevant in determining the offensive capability of a given Pokémon. In the game,
“Each Pokémon has six major Stats, which are as follows: HP, Attack, Defense, Special Attack, Special Defense and Speed. HP means ‘Hit Points’ and represents health (‘amount of vitality’) of a Pokémon. When it suffers damage, a numerical value is calculated by the game, and the result is subtracted from the current HP. When HP reaches zero, the Pokémon faints and is out of action.”
― Vianna Sym (2015: p. 26), my translation
In the games, Pokémon are defined by certain features, among which are the above-mentioned Stats. Each Pokémon has a given number of points assigned differently to its Stats, making it tough, agile or strong. HP represents the Health Points (or Hit Points) of a Pokémon, and from the work of Codd (2016), it is understood that a Pokémon that is “powerful” is the one with the highest chances to take the opponent’s HP down to 0 more effectively.
Thus, to estimate how powerful a Pokémon is, one should not base his/her calculations on features estimated from the anime, but rather analyze the Stats distribution of a given Pokémon as it appears in the game. This study takes into account the Stats of each of the starting Pokémon to more thoroughly analyze how powerful each can become, that is, how much damage a Pokémon can cause in a battle.
Let’s first set the game to be any of the so-called “Gen I” versions (Pokémon Red, Blue, Green or Yellow), released between 1996 and 1998. In these versions of the game, there were less Stats, only: HP, Attack, Defense, Special and Speed (also, there were no mega-evolutions). The distribution of stats between the starting Pokémon (in their last form) can be seen in Figure 1.
Figure 1. Base stats of (from top to bottom) Venusaur, Charizard and Blastoise in Gen I. Source of the tables: Serebii.net. Original artwork of the Pokémon by Ken Sugimori; available through Bulbapedia.
By comparing the so-called Base Stats of the three starting Pokémon (from Fig. 1), we get the chart shown in Figure 2. This gives us a broader view of the Stats distribution of each Pokémon, distinguishing their higher and lower attributes. If we add up all the Base Stats of each Pokémon, we obtain a grand total score of Stats points (Fig. 3). From Figure 3, it can be seen that all three Pokémon sum up to the same value: 425 points. In the first versions of the games the Stats were kept in a balance during the development of these three Pokémon. Thus, the sum of Base Stats alone is not enough to show which starter is the strongest. There’s more to consider.
Figure 2. Chart comparing the Base Stats of the three starters in Gen I.
Figure 3. Sum of all Base Stats values of each starter Pokémon in its final form (Gen I).
Using the Base Stats, we can estimate the possible amount of damage (measured in hit points, or HP; Vianna Sym, 2015) that a Pokémon can cause with one of his moves. This is in fact based on a complex calculation depending on several variables, such as the attacking Pokémon’s level and offensive Stat and the opponent’s defensive Stat, alongside some occasional bonuses. By default, the formula is expressed as (Vianna Sym, 2015):
where “Level” is the current character level of the attacking Pokémon, ranging between 1 and 100; “AttackStat” is the Base Attack Stat or Special Stat (depending on the kind of move, Physical or Special, used) of the attacking Pokémon; “DefenseStat” is the Base Defense Stat or Special Stat (again, depending on the kind of move used) of the opponent; “AttackPower” is the power of the move used (this is pre-defined in the game and each move has its own power value), where a greater value represents a greater damage output; “STAB” is an acronym for “Same-Type Attack Bonus”, which means that if the move used has the same type as that of the Pokémon using it, it increases in 50% (STAB = 1.5; otherwise, STAB = 1); “Weakness” is applied depending on whether the chosen move is super effective on the opponent (this variable can assume values of 0.25, 0.5, 1, 2 or 4, depending on the type of the move and of the defending Pokémon); “RandomNumber” is simply an integer assigned randomly by the game, ranging from 85 to 100.
Other in-game factors may cause changes in damage output, for example: weather effects (rain and sunshine), and the so-called “buffs” and “de-buffs”, which are respectively temporary increases and decreases in the Pokémon’s Stats caused by moves such as Agility, Dragon Dance, Swords Dance etc. Weather effects were not yet present in the first versions of the game, so they will not be considered in this study. Moreover, to keep the analysis simple (not to say feasible), increases/decreases in Stats will also not be taken into account. The calculations here use only the Base Stats of the Pokémon in question and the set Power value of the moves. Weakness will also not be applied.
Codd (2016) considered the “signature moves” of the starting Pokémon as: Solar Beam for Venusaur (grass type), Flamethrower for Charizard (fire type), and Hydro Pump for Blastoise (water type).
The Power of each of these moves can be seen in Figure 4, alongside other data: “Battle Type” is the type of the moves, which in this case are the same as the types of the starter Pokémon (so STAB = 1.5); “Category” refers to whether the move is a Physical Attack or a Special Attack (all are Special and thus use the Base Special Stat); “Power Points” (PP) represent the number of times that the move can be used; “Power Base” is the Power of the move (used in the equation above); “Accuracy” refers to the probability of success in hitting the opponent (in %).
Figure 4. From top to bottom, the moves Solar Beam (formerly rendered as “Solarbeam” or “SolarBeam”), Flamethrower and Hydro Pump, showing their in-game Power values and type (in Gen I). The symbol in the “Category” entry means that the moves are all Special Attacks. Source: Serebii.net.
CALCULATING THE DAMAGE
To calculate the damage dealt by each of the starter Pokémon with their signature moves, I used a virtual calculator available at Smogon University, the “Pokémon Showdown”. (Smogon University is a community dedicated to the competitive world of Pokémon games, giving the players some useful tools.) The moves have the Power values shown in Figure 4 and the defending Pokémon will be a Chansey (see Fig. 5 for Base Stats), which is neutral (that means, neither weak nor strong) towards the starters and their signature moves. All Pokémon are considered to be Level 100.
Figure 5. Base stats of Chansey in Gen I. Source of the table: Serebii.net. Original artwork of the Pokémon by Ken Sugimori; available through Bulbapedia.
By putting all the values in the Pokémon Showdown calculator, we have:
Venusaur (Solar Beam): Note that the Gen I version of Solar Beam is not present in the Pokémon Showdown database, so I used the Gen II version instead (the Power is the same). The damage output falls in the interval 125 to 147 points, which represents 17 to 20% of Chansey’s total HP. Venusaur needs to land 5 blows to knock out its target.
Charizard (Flamethrower): The damage output falls in the interval 90 to 106 points, which represents 12 to 15% of Chansey’s total HP. Charizard needs to land 7 blows to knock out its target.
Blastoise (Hydro Pump): The damage output falls in the interval 113 to 133 points, which represents 16 to 18% of Chansey’s total HP. Blastoise needs to land 6 blows to knock out its target.
Just in case, these numbers were checked on another calculator, built by myself (Pokémon Damage Calculator; Carli, 2016). An algorithm was developed based on the damage equation from above, translated in some programming languages (available at: https://github.com/brunolcarli/pokeDamageCalc) and then translated into APK format so it can be installed on any mobile device running on Android (Fig. 6) or Windows operating systems. Feel free to download the app at: https://build.phonegap.com/apps/1824036/install. The results were very similar (Fig. 6): 127 to 144 points of damage for Venusaur’s Solar Beam; 84 to 98 points of damage for Charizard’s Flamethrower; and 106 to 122 points of damage for Blastoise’s Hydro Pump.
Figure 6. Screenshots of the Pokémon Damage Calculator app (Carli, 2016: v. 1.0.0, running on Android OS), showing the maximum damage output for Venusaur’s Solar Beam (left), Charizard’s Flamethrower (middle) and Blastoise’s Hydro Pump (right).
Organizing all these numbers (from both the Pokémon Showdown and the Pokémon Damage Calculator) in a chart (Fig. 7), it is possible to clearly see the minimum and maximum damage each of the initial Pokémon can inflict, with their signature moves, against a neutral target. It can be seen that Charizard is actually the Pokémon that causes the least amount of damage, while Venusaur can deal the greatest amount of damage. Thus, Venusaur can be regarded as the “most potent” starter if we are referring to the sheer amount of damage caused.
The present study thus shows that Codd’s (2016) analysis is not applicable to the game itself, since it is not based on the variables and values present in the game mechanics. Also, as shown above, Venusaur and not Charizard is the “most potent” starter considering just the raw amount of damage it can cause. However, this is true only for a single attack in a single round of battle (which is important for the so-called “one-hit knockout”). Of course, as every player knows, one should not think that damage output alone makes a Pokémon more effective in battle. The game has much greater complexity and we would be reducing it to nothing if we just consider maximum damage. For instance, Solar Beam is a move that needs to spend 1 turn of the battle recharging, while both Flamethrower and Hydro Pump can be used every round. Furthermore, there are other factors, like Hydro Pump having an accuracy of 80% (meaning it misses one out of every five times) and Flamethrower being able to leave the defending Pokémon with the burn status condition. However, this is a matter for another day; for now, Charizard has lost its crown.
Figure 7. Simple chart showing the maximum (red) and minimum (blue) points of damage each of the starters can inflict with their signature moves (Solar Beam for Venusaur, Flamethrower for Charizard, and Hydro Pump for Blastoise). The chart takes into account the values obtained by both the Pokémon Showdown and the Pokémon Damage Calculator.
As any normal 10-year-old, I chose Charmander as my starter in Pokémon Blue. I would demolish any and every Pokémon that came after me and my little (pseudo-)dragon, but Blue’s (or Green’s, depending on which game you were playing) Squirtle and its evolutions would always put a big dent on my flaming lizard’s health, even KO’ing him sometimes. Losing my beloved Charizard would leave me with an undertrained Pidgey and some other HM slaves, so Blastoise did leave a sour taste in my mouth growing up. As a kid, I had a hard time understanding how that big fat turtle/tortoise could hurt so much my giant fire-breathing monster, but now, as a bigger kid, I decided to take a more detailed look at Blastoise’s power.
HYDRO PUMP ESTIMATION
Bulbapedia (2015) states that Blastoise is 1.6 m tall and, from images from the anime, its shell was determined to be of ellipsoidal shape, with axes of around 1.32 m, 1.33 m and 1.42 m (Figs. 1–3). Its volume can then be calculated as follows:
Where a, b and c are the lengths of the axes, which yields an internal volume of 1.305 m³. If the carapace were filled half with organs and muscles and half with water, there would be approximately 653 liters of liquid stored in it. The cannon’s diameter was assumed to be around 10 cm each.
Figure 1. Blastoise, as seen on the anime. Image modified from a screenshot.
To calculate how strong its water attack is, a Hydro Pump (which has 5 Power Points [PPs], meaning that the Blastoise can use Hydro Pump 5 times) was compared to a water cannon. Information about water cannons can be found in a study made by The Omega Research Foundation (2000), which is an independent UK-based research organization (Fig. 4). This paper states that low pressure water jets have a pressure of about 150 psi (1.03 MPa), and some modern ones can have as much as 360 psi (2.48 MPa). As Bainbridge (2014) says, Pokémons are supernatural creatures that possess spiritual and supernatural powers (even Blastoises, I guess), so it will be assumed that the water in the shell is stored at 360 psi, the higher value.
Figure 2. Blastoise sideways. Image modified fromtheofficial art by Ken Sugimori. (Source: Bulbapedia.)
As the cannons are short tubes and the only losses come from its entrances and exits, which amount to very little, the losses will be considered negligible. The time needed for the water to accelerate will also be considered negligible, since there is a big pressure gradient and no obstacles to the flow. This way, all this pressure would translate to momentum when the water left the two cannons, with the velocity being calculated as:
In this equation, c is the velocity of the water that exits the cannon, p is the pressure at which the water is stored and is the density of water. The equation is used to transform the stored pressure cargo into velocity when there is no vertical movement in the flow.
Figure 3. The “Blastoise ellipsoid”.
At 360 psi and the water density of 1000 kg/m³, the exit speed would be of 70.5 m/s, an astounding 254 km/h. However, all that might would be very short lived: if all the stored water is to be spent only in Hydro Pumps, each PP would last only 0.118 seconds, so Blastoises must be very good shots (no time to redirect the attack, which would explain the low accuracy for Hydro Pumps). Its force and energy were calculated as follows:
Figure 4. Modern Israeli pulsed jet Water Cannon. Image reproduced from: The Omega Research Foundation (2009).
The first equation is the force exerted by the flow, which is given by the multiplication of minus its mass flow (negative, as the water is flowing out of Blastoise) by the velocity of the water. The second equation is the kinetic energy of the moving water, given by its volume, density and speed squared. That translates to a 78 kN impact for 0.118 seconds, and about 324 kJ of energy. While it does not pack as much force as Ivan Drago’s punch (about 130 kN according to Rocky 4; his “punching power” was shown as 1850 psi; the punch’s area estimated at 0.01 m², the approximate area of a closed human fist), energy wise, one PP would have the same energy as an average car (1300 kg) going at 80 km/h. That would be tough for a Charizard to handle.
So, after all, my Charizard really should avoid getting hit by one of those things. Some follow up questions do come to mind though, so let’s see what else we can come up with.
IS MEGA BLASTOISE’S HYDRO PUMP STRONGER?
For this analysis, let’s assume that the pressure at which the water is stored in a Blastoise’s shell is directly proportional to its Special Attack. Blastoise’s base Sp. Attack is 85, whereas its Mega Evolution packs 135 (a regular Charizard has 109 base Sp. Attack, while its Y Mega Evolution has 159, so take that Blastoise!). The water pressure in Mega Blastoise’s carapace would be of 571.8 psi (3.94 MPa). Another difference is in the cannon: while Blastoise has two 10 cm diameter cannons, Mega Blastoise has one that is significantly larger, assumed to be 20 cm in diameter, as seen in Figure 5.
Figure 5. Comparison between Blastoise and Mega Blastoise. Image modified fromtheofficial art by Ken Sugimori. (Source: Bulbapedia.)
With these assumptions set, the force of a Hydro Pump’s PP would be of almost 250 kN, way stronger than Ivan Drago’s punch, and the energy it would contain would be of 515 kJ. Overall, Mega Blastoise really is stronger than its non-Mega counterpart, with its Hydro Pump carrying approximately 60% more energy and exerting a force 3 times stronger.
CAN A BLASTOISE FLY?
With that amount of force and energy being released, one wonders if a Blastoise could fly (Doduos can, why can’t the turtle?). In order to calculate how high can a Blastoise go, the Pokémon was supposed to be in an upside down position and to fire its Hydro Pump downward, with the water jet now propelling the turtle upward. This maneuver was approximated to a simple variable mass system, and calculations without considering drag forces were, as stated in Peraire & Widnall (2009):
The first equation calculates how high a body with mass goes when continuously firing downward a total mass of propellant at a given velocity c and at a mass flow while being accelerated by a gravity (9.8 m/s²). The auxiliary variables and n are the mass fraction, which is the quotient of the final mass of the body after the propellant is expelled divided by its initial mass, and the thrust induced acceleration. In this study, the body is a Blastoise and the propellant is water.
The results are shown in Figure 6, a graph representing how high Blastoise goes according to how many PPs it uses.
Figure 6. Graph of achieved height versus number of PPs used for Hydro Pump.
Therefore, of course, the more Hydro Pumps used, the higher the Blastoise will go. When all five PPs are spent, Blastoise will fly as high as 28 m, which is about the height of a 10-story building. That will not get you from Pallet to Viridian, but it is pretty high for a turtle.
HOW MUCH MUST A BLASTOISE EAT?
It was hard to find information on the efficiency of turtles, so a human was taken as basis of comparison. An average human has a mechanical efficiency of about 25%. As turtles are ectothermic creatures, energy-wise they are more efficient than humans, as they do not spend energy to stay warm. Its efficiency is then a little higher and, as a semi-wild guess, it was assumed to be 40%.
To most people’s astonishment, freshwater turtles are not herbivores. They do eat plants, but fish is a main dish on their diet. Blastoises, as freshwater turtles, would then eat what should be easier for them to find: Goldeens. Goldeens weigh about 15 kg and inhabit the same water ponds a Blastoise would. Nutritionally, freshwater fishes have about 150 kcal per 100 grams, so with 40% efficiency a Blastoise would have to eat a little more than half a kg of Goldeen to use all 5 PPs of Hydro Pump, which is a little underwhelming, as we were expecting some mass extinctions to take place.
After our analysis of a Blastoise, it was concluded that they do pack a punch. With each Hydro Pump averaging 78 kN in force and 324 kJ in energy, the impact could damage any fire lizards that stands on its way. Aside from that, Mega Blastoise was confirmed to be stronger than a regular one, and Blastoises do not need to go binge-eating Goldeens to use their attacks. The most important result, however, was that turtles, even when equipped with super strong water cannons, in fact cannot fly.
Bainbridge, J. (2014) “Gotta Catch ‘Em All!” – Pokémon, Cultural Practice and Object Networks. The IAFOR Journal of Asian Studies 1(1): 1–15.
Since Pokémon is a recurrent topic on this journal, I would like to call your attention to this little fellow: the fletchling.
Fletchling (yayakoma, in Japanese), as it appears in official Pokémon artwork.
Fletchling is a tiny normal/flying-type robin pokémon with an orange head and grey body. Both Pokédex and Bulbapedia tell us that they sing beautifully, send signs using chirps and tail movements and are also merciless to intruders in their territory. It evolves to a fire/flying peregrine falcon (how a robin becomes a falcon is a topic for further discussion) that is a very common sight in competitions.
Back to fletchling. Even though I like all sorts of birds (I am an ornithologist after all), we always have our favorites; mine is the robin. And so, the tiny robin fletchling became my all-time-favorite pokémon. Now let us take a look at the robin I find in my garden.
Well, they look somewhat similar, but the color differs. Could my garden robin and fletchling be the same thing then? Are there any other robins outta there?
No no, I meant bird robins.
So it is finally clear that fletchling was based on the Japanese robin and not on the European one from my garden (even though the entire Pokémon XY games supposedly been based on France – good job, Game Freak Inc.).
Now let us take a closer look at the bird robins (please refer to the figures above). We can see that the Japanese and European robins are very similar between themselves, especially when you compare them to the American and Australian robins (see figure below). This is expected, since the former share the same genus (Erithacus), meaning that they are more closely related. That is why they are so similar in appearance despite the difference in color. There is yet another Erithacus robin in Japan which has even more distinct plumage color (the Ryukyu robin, see figure below), but that is still very similar in shape to the European and Japanese robins.
American robins, on the other hand, are much more different. They belong to another genus (Turdus), which also includes blackbirds, song thrushes and fieldfares. As such, they are only distantly related to the species belonging to Erithacus. Actually, Turdus might even belong to a completely different family – this is a hotly debated topic in ornithological circles, but I will not dwell on it.
So why we call all these different birds “robins”?
Robin is a popular English name to refer to passerines with red breast. The first one to be named as such was the European robin and the name was later on “exported” by colonizers and travelers for the birds in other continents. In other languages, the red breast feature of the European robin is always the focus: “Rotkehlchen” (German), “pisco-de-peito-ruivo” (Portuguese), “rouge-gorge familier” (French), “petirrojo” (Spanish), “pettirosso” (Italian) etc. Folklore says the red breast was earned by the brave small European robin as a token for its heroic acts (Greenoak, 1997).
European/Japanese and American/ Australian robins all share the red breast feature, being, thus, all called “robins”. However, as we saw, one pair is not closely related to the other – they do not share the same genus. This is because their popular name is not based on any evidence of how closely related they are. Popular names are just useful tools for people’s everyday life. Scientific names, however, are more than that. As we saw, color is not the only characteristic that make a bird a Turdus or an Erithacus – The other Japanese robin (the Ryukyu robin) does not even have an orange breast; what makes it an Erithacus is its body shape, skeleton, anatomy etc. Giving a name is not an easy matter in science (this branch of Biology is called Taxonomy, by the way). The act of classifying and naming a species is based on studies that analyze the morphology and even the DNA of living beings to decide who is more related to whom (and therefore belong to the same genus or family). Therefore, scientific names also contains information on the relationship between species and will never confuse someone as popular names like “robin” do.