ID 436: BellossomBellossom is more or less one of the most memorable designs created for Gold and Silver; to this day, I think that next to the Eevees, Pichu, and maybe Hoothoot, more casual Pokemon fans would recognize Bellossom than any other designs from Generation II. Clearly, the team did something right with this cute hula-dancer. So let’s take a look at what went into Bellossom. As I’ve mentioned many times by this point, Period 3b of development had a very clear mission: take Generation I Pokemon and extend their evolutionary family in some way. We’ve seen a number of different ways to do that: Steelix, Plux, and Scizor were trade-only evolutions that added the Steel-type, giving each of these a unique identity. Madame and Porygon2 both took mostly unused Pokemon with middling stats from Generation I and buffed their stats up. Beyond stats, Madame and Porygon2 were more or less identical to their progenitors. And then there’s a whole bunch of baby Pokemon which tried to extend a Generation I family by creating a weaker and cuter version of the original. These babies could potentially be encountered earlier, allowing mid-game Gen I Pokemon to gain a new life as an early game encounter. The main thing that’s missing from this section, so far, is branching evolutions. By branching evolutions, I mean what we've previously seen with Politoad and Slowking: split evolutions in which the player has a choice to evolve the second form into on of two unique final forms. To some degree, the Eeveelutions in the three previous entries we discussed are examples of branching evolutions, but given that Eevee’s distinction has always been that it evolves into multiple forms, I don’t think this quite counts. Still, I don’t think its coincidence that Bellossom and Tsubomitto directly follow from the three Eevees—I think it’s quite likely that, reminded of the concept of branching evolutions by creating the Eevees, that same creator (probably Nishida) decided to fill the gap in Period 3b by designing two more branching evolutions. We’ve seen branching evolutions in the Korean Index before, but it was a long time ago: Politoad was created by either Nishida or Sugimori in Period 1c and Slowking was in Period 1d. These designs were probably at least a year to a year and a half old by this point in development, and it’s a bit strange the team hadn’t followed up on the idea of branching evolutions since then: it’s a pretty cool concept and a natural one for a sequel. But given Era III’s mandate to round out ideas created earlier in development, it’s natural to find two more branching evolutions right before 1997’s development was completed. Naturally, the development team focused on two parallel families from Generation I: The Bellsprout family, and the Oddish family. In both Bellossom's case, and it's Bellsprout counterpart Tsubomitto, both alternative evolutions are far more humanoid than their alternatives, Vileplume and Victreebel. In Bellossom's case, the team decided on a female appearance and made Oddish's leaves sprout and become flowers on its head! It also grew more leaves around its hips, which resemble a Hula-Skirt. And Bellossom was born! The Oddish line (and as we’ll talk about, the Bellsprout line) were good candidates for a new branching evolution: since they already evolved via evolutionary stone, the developers could pretend that they always had the ability to branch into a separate evolution, using an evolutionary stone that simply wasn't available in the first games. In the case of Bellossom (and Tsubomitto, the Bellsprout branching evolution), the decision was made to have them evolve via the newly introduced Poison Stone, which wouldn't have been possible in Generation I. As a quick aside, the Oddish family and the Bellsprout family have an interesting history in Generation I. They were the last Pokemon designed for those games, excluding Mew. As far as we can tell, when the choice was made to change the central elemental triad in Generation I from Fire/Water/Electric to Fire/Water/Grass, the team realized they needed more Grass Pokemon to give the players more choices, especially if they chose a starter other than Bulbasaur. Atsuko Nishida seems to have designed both families at the last second to fill this gameplay need. It's also interesting because these two families were potentially the only ones created with the idea of version exclusives in mind. While the idea of different games that had different stories or Pokemon was around since the very beginning of development, the idea of two different games with Pokemon that were exclusive to one or the other game probably only came together very late in development. Oddish and Bellsprout may have been designed as version exclusives, while the rest of the exclusives were pieced together at the last second. Anyway, given their places at the very end of the Korean Index, that Bellossom and it's counterpart Tsubomitto were designed close to the end of Spaceworld '97's development as well. In fact, by the time of the Spaceworld ’97 build, Bellossom was clearly only a vague idea. In the demo build, it's identical to Vileplume. Bellossom has the same typing as Vileplume—they’re both Poison/Grass—and they both learn the exact same four moves, all at level one: Absorb, Stun Spore, Acid, and Petal Dance. Almost none of the new Pokemon have stats in Spaceworld ’97, and Bellossom was typical, with only placeholder stats that don't suggest any differences from Vileplume. The only difference between them was that Bellossom evolved with the Poison Stone (which didn’t yet work) and Vileplume used the Leaf Stone. At this stage of development, the team either hadn’t figured out a way to distinguish Bellossom as a new and interesting choice for a Pokemon trainer to make, or they hadn’t had time to implement their ideas. Which makes sense. Given how late we are into the Korean Index, Bellossom was probably only added to the Pokedex months--maybe even weeks--before the Spaceworld 1997 build, and it isn’t a surprise the team just hadn’t gotten around to it yet. Tsubomitto has the exact same problem that Bellossom does, more proof that the ending of the Korean Index was finalized right before the Spaceworld ’97. Another part of Bellossom's problem, though, was more fundamental. In Generation I, most Pokemon which evolved through an evolutionary stone didn’t learn any new moves after they evolved (minus a few exceptions, like Exeggcutor learning Stomp because Exeggcute clearly couldn’t). Vileplume and Victreebel technically do, but I'll talk about that technicality in the next entry; suffice to say, it isn't meaningful. The reason stone evolutions were designed this way was to create tension between evolving the Pokemon the first chance you get, or waiting until your Gloom learns the right moveset before you get the raw power of Vileplume. Evolve Vileplume too early, and you might miss out on some of its best moves. I know the first time I played the game I went through the entire game with a Clefable I evolved instantly: as a kid I had no idea this was the wrong thing to do, and that poor Clefable only knew Pound and never got Metronome. This is a good balancing conceit, but it creates an obvious problem when you want to design new branching evolutions. If neither Bellossom nor Vileplume can learn any different moves after evolving, then how are you supposed to make them different from each other? There’s only two strategies: change their typing, or change their stats. However, a change in typing is difficult to pull off if the Pokemon can’t learn new moves: a Generation I Poliwrath can’t learn fighting moves except by TM, for instance, making its new type more a detriment than anything else; even with its stronger stats, most players of Gen I opt to keep Poliwhirl because Psychic isn't super effective against it. The second option, giving the two counterparts different stats, is fine, but pretty subtle to casual players who might not encounter both evolutionary forms and not really know the differences. The development team made a minor change to these rules about movesets by the final release of Gold and Silver, giving some stone evolutions a few moves to make them unique. Poliwrath, for instance, now learned Submission and Mindreader despite being a stone evolution (though poor Clefable was still moveless). In the Oddish family’s case, Bellossom at least got to learn Solar Beam at level 55, to differentiate her from Vileplume and Gloom, which cannot learn that move. The team slightly differentiated Bellossom’s stats from Vileplume—Bellossom has 100 Special Defense and 90 Special Attack, while Vileplume flip-flops them—but surprisingly, they let most of the stats identical, and even these differences over Special are incredibly minor. It’s honestly bizarre to me that even with all that extra development time, Bellossom and Vileplume ended up so similar to each other. The main way they differentiated the two was by making Bellossom pure-Grass type rather than Grass/Poison, like Vileplume and the rest of the Oddishes. In retrospect, this was a good idea. For reasons that are unfathomable to me (but are almost assuredly due to some quirk of how the Grass type was originally invented in Gen I development), almost every single Grass type in those games is dual typed with Poison, and the Oddish line is no exception. In Gold and Silver, there are some modest experiments with Grass types that aren’t Poison-type: the Hoppip line is Grass/Flying, and the Sunflora family is pure Grass (Grass/Psychic in SW’97). By letting Bellossom lose the Poison-type, it at least switched up some of the match-ups for the Pokemon, giving the player a reason to choose it over Vileplume. They couldn’t have done this in Spaceworld ’97, for a simple reason: Bellossom evolved via the Poison stone. But that stone got scrapped and the team replaced it with the Sun Stone, which was used to evolve both Sunkern to Sunflora and Gloom to Bellossom in the final games. So you were free to make a choice between Vileplume and his new counterpart. EXCEPT—And I didn’t quite realize this when I was discussing Sunflora—the developers of Gold and Silver made another incomprehensible decision that make Sunflora and Bellossom more or less impossible to use in a regular run. Because there’s only one way to get a Sun Stone: win the Bug Catching Contest. Which you’d think is easy, but the way it is scored, is more or less nearly impossible. In the games, another one of the competitors is a blue haired trainer named Cool Trainer Nick. He always wins: he gets a minimum of 349 points in the contest, which is almost the maximum you could get with a perfect IV, full-life Scyther, which would have 386 points. Thus, unless you win the lottery, Cool Trainer Nick will always beat you in the Bug Catching contest, making the Sun Stone more or less out of reach for any casual player. As a result, without gaming the system and knowing exactly what you’re doing, no normal player is ever going to get a Sun Stone. And thus Bellossom, like so many other new Gen II Pokemon, is basically off limits for the player. Let’s also talk about Bellossom’s looks. I have a bone to pick with Bellossom. I’ve always disliked how these branching evolutions look nothing like the rest of their family. I mean, c’mon: Poliwag, Poliwhirl, and Poliwrath all have a design consistency: they share the same palette, the same whirl on their belly. And then Politoed is a completely different color, completely different sillouette, completely different body shape. Bellossom, likewise, looks nothing like the rest of the Oddish family and it will continually annoy me. (Of course, we have two reasons for why Politoed looks so different from the rest of its family. First, according to interviews, the design team heard a lot of criticism from Japanese players who wanted the tadpole Poliwhirl to grow up into a frog, so they made Politoed actually become a frog. Secondly, as I commented before, there’s good reason to think that Politoed was designed from the alternative cut frog line from Generation I, and probably inherits some of its dissimilar looks from that origin). Bellossom is even more frustrating than Politoed. The final Bellossom looks absolutely nothing like the Oddish line. It’s the wrong color, it is shaped differently, it doesn’t drool like Gloom and the flowers on its head look nothing like the Rafflesia flower that Gloom has (okay, okay, they've got five petals and the same center as Vileplume's, but they're much smaller, the wrong color, and lack the dots Rafflesia are known for). It has none of the defining features of the Oddish family. Honestly, as a kid I thought Bellossom was a completely new Pokemon and I was amazed it evolved from Gloom. I mean, do these look anything alike? What makes this so frustrating is that we now know it wasn’t always the case! Early Bellossom looks obviously like an Oddish evolution! Even the palette they had chosen for Bellossom by June 1999 obviously looks like an Oddish evolution! It's not perfectly in line with the rest--in the SW97 sprites Bellossom's flowers are a different color, and in the June 1999 sprites, its skin is a lighter shade of blue compared to the rest. But both clearly at least look like part of the Oddish family! Okay okay, so obviously the Spaceworld '97 sprite isn’t great, and needed to change. Given that during development the team became aware of the racist implications of Jynx’s design, a black-faced hula dancer was…probably not the right direction to go. It’s good that the team continued to iterate on this original design. I'll champion them for dropping this pretty racist sprite and trying a different direction. I will say one thing about that Spaceworld '97 sprite: I really like how it has a continuity with the "scent particles" that surround Gloom. Unlike Gloom's smelly gook surrounding it though, early Bellossom seems surrounded by sparkles, maybe suggesting that it smells much better than Gloom. (Thanks to @OrangeFrench for pointing this detail out!) The team tried changing the color of Bellossom to a light blue in by the time the Gameboy Color had given them a much wider palette to play with, and this looks a lot, well, less racist. It’s not quite the same color as the rest of the Oddish family, but you can still easily see the family resemblance. Moreover, with this palette, Bellossom’s leaves were red, which made it more in line with Vileplume’s red flower. I like this palette; it solves the racism problems but still makes Bellossom look like the rest of the family. And it appears that the team was happy to go with this color scheme for an awful long time. Sugimori’s early artwork of Bellossom uses this color scheme, for instance. But by Spaceworld ’99 (and probably right before that build, given how Sugimori’s artwork still used the old palette) someone chickened out. While nothing’s confirmed, I think it’s safe to say that someone in Game Freak (maybe a Nintendo executive, still very sensitive to the Jynx controversy) worried that the blue skin was still too close to blackface. By Spaceworld ’99, the blue color scheme became Bellossom’s shiny color, but normal Bellossom had been changed to a light green. Crystal, in turn, cemented this change by giving Bellossom and even darker shade of green as its skin color. I hate this. I'm fully in favor of fixing Pokemon designs to avoid offense. That's not the problem here. But the green color they chose ended up ruining Bellossom's design aesthetic completely. While if you trace the development of Bellossom from the beginning you can see its resemblance to Oddish, viewed in isolation it makes no sense; it’s completely at odds with the rest of the family. Combine this terrible choice in palette with the virtual impossibility of a player using Bellossom, and it feels to me like Bellossom was just a real missed opportunity in Gold and Silver. It’s a good idea, and the anime certainly popularized Bellossom enough to make it mildly iconic. However, the choices Game Freak made concerning this Pokemon were such obvious mistakes in retrospect. It’s a real disrespect to Bellossom, a Pokemon that could have been so interesting, and turned out to be an inaccessible, badly colored, near clone of Vileplume. ID 436: TsubomittoThe final mistake the designers made with Bellossom was giving it a spot in the final games but cutting its much-superior counterpart, Tsubomitto. Tsubomitto’s great! If Bellossom is the happy-go-lucky goody two shoes of the Oddish family, then Tsubomitto is the bad boy Bellsprout that doesn’t take shit from anyone. Don’t run into this guy in a blind alley, or you might not make it out. Tsubomitto is, obviously, an alternative evolution for Weepinbell, competing for Victreebel’s place in the family line. Victreebel takes after Weepinbell, accentuating the deadly mouth it has and foregoing legs and arms. Tsubomitto, on the other hand, takes after Bellsprout and has a full two legged body: unlike Bellsprout’s tiny wisps of roots for legs, Tsubomitto has chunky legs that may be tubers repurposed for movement. It's unclear exactly where Tsubomitto’s name comes from. “Tsubo” appears in Bellsprout’s Japanese name (and sort of in Victreebell’s, if you mash two words together). In Bellsprout’s name, it’s part of the word “Tsubomi,” or “to bud.” Maybe it’s a reference to the same verb in Tsubomitto’s name, or maybe the "Tsubo" means “pot,” which is what the word can mean without the “mi” attached. “Pot” in its name could designate the pitcher plant part of its face; I don’t really know. “Mitto” is probably a transliteration of the English word “Mitt,” like a catcher’s mitt; given that the Bellsprout line are pitcher plants, a pitcher holding a mitt would be a funny pun. I’m not sure where’s Tsubomitto’s mitt is actually located, but this origin seems plausible to me. While we're talking about Tsubomitto's sprite, can I also just point out that I love the way it seems to have a superhero cape made out of one of its leaves? In the Spaceworld '97 version of the sprite, the large leaves Tsubomitto has on its sides are its hands, but in the revised one seen above, the artist added smaller root hands under the leaves, making the leaf around its neck look majestic or part of Tsubomitto's superhero identity. Another reason it and Madame are the two coolest lost Pokemon. Like I said in the previous entry, Tsubomitto is more or less identical to Victreebel in Spaceworld ’97. Like Bellossom’s relationship with Vileplume, Tsubomitto learns the exact same moves as Victreebel, it has the same typing as Victreebel, and it has placeholder stats. The only difference is that it evolves (like Bellossom) using the newly created Poison Stone, which doesn’t work in the Spaceworld ’97 demo. Again, I think the stark similarities between Tsubomitto and Victreebel might be an indication of just how close the end of the Korean Index was to the Spaceworld ’97 build; it could also just be the case that it’s difficult to differentiate two branching evolutions when the rules the team was using said that stone evolutions typically didn’t learn many new moves (if any at all). Quick note about the moveset, by the way. I mentioned in the previous entry that Victreebel and Vileplume only learn four moves at level one, exactly mimicked by Bellossom and Tsubomitto. Which is true for Generation II, but apparently in Generation I, each of them did learn three level up moves, though these level up moves were all learned before Bellsprout and Oddish evolved into Weepinbell and Gloom, thus before any player could legally obtain either Victreebel or Vileplume. For instance, while in Generation II Victreebel only learns Vine Whip, Sweet Scent, Sleep Powder, and Razor Leaf, in Generation I it also learned Wrap at level 13, Poison Powder at level 15, and Sleep Powder at level 19. You've never know this though, since a Bellsprout doesn't evolve in Weepinbell until level 21, by which point it'd be too late to learn those moves. Those moves are probably there so that Erica’s Pokemon have a slightly better moveset, but otherwise no player would ever notice these level up moves exist at all. Strangely, these level up moves were completely taken away from Vileplume and Victreebel in Generation II, though I don’t know why any developer would both making a change to something so minor as to be unnoticeable in regular gameplay. Tsubomitto did make it through the 1998 development hiatus, unlike so many other SW’97 Pokemon. But by the time we next have a snapshot of it, Tsubomitto’s gone through some changes. Most noticeably, its got a new fabulous striped look, with new hands and a cool leaf-cape. Secondly, by April 1999, the team actually removed Tsubomitto’s evolutionary method, so that Weepinbell doesn’t evolve in it at all! Tsubomitto will eventually evolve via the Sun Stone, just like Bellossom, but not until three months later, at the end of July. This, to my mind, suggests two possibilities. First, the team briefly planned to removed Tsubomitto, and the development team started the process by removing its evolutionary method before changing their minds. Secondly, it makes me wonder if Tsubomitto was rebuilt from the ground up in the 1999 reboot, and the team just hadn’t bothered programming in an evolutionary method. It could also be simpler than that: maybe they just hadn't programmed the Sun Stone into the game by April, but they had cut the Poison Stone, so there was briefly no option for Tsubomitto. (This section was updated to reflect info provided by GOLDS_TCRF in the comments! Thanks!) Let’s also put the history of Tsubomitto’s moveset back-to-back with this. In April 1999, when it didn’t evolve from anything, Tsubomitto had a slightly different moveset: strangely, it had Victreebel’s moveset from Generation I, which meant that it learned the same four moves at level one as it always did… plus Sleep Powder, Poison Powder, and Wrap at a level too low for a player to ever encounter. Weirdly, Victreebel still doesn’t have these moves, so for some reason Tsubomitto’s moveset was pulled from Generation I, not copied from Victreebel like it was in Spaceworld ’97: some evidence that, for whatever reason, Tsubomitto was rebuilt from scratch in April 1999. Maybe the team had planned on deleting Tsubomitto initially when they were rebuilding the game in 1999, but changed their minds when they had an extra slot Tsubomitto could fit into? By July 1999, it’s got an entirely full moveset, though it’s just a copy of Weepinbell’s, with a couple of notable changes. Instead of Slam, Tsubomitto learned Solar Beam (maybe a change made because it now evolves via Sun Stone) and instead of Sweet Scent, it learned Lovely Kiss, Jynx’s signature move. Which is honestly wonderful: the image of Tsubomitto suctioning an enemy Pokemon with that face is priceless, and it gives Tsubomitto some badly needed character. This is still a weird moveset: while it would work to let Weepinbell choose when to level up and thus choose which moves it got, that’s normally not a pattern we find in Stone evolutions. The team agreed, because by August 16th, it’s back down to four level one moves and only Solar Beam as a level up move. The Solar Beam is learned just a level before Bellossom’s, and it now has Lovely Kiss as one of its level one moves. On the one hand, great: I'm glad they kept this flavorful move on Tsubomitto. On the other hand, boo: if the move is learned at level one, there would be no way for the player to ever get it! Lovely Kiss would appear purely on enemy Tsubomitto. Definitely an odd choice that still feels unfinalized. As the team was figuring out the moveset, they were also trying to make Tsubomitto feel distinct from Victreebel by differentiating the stats of the two. By April 1999, Tsubomitto no longer had placeholder stats; instead, it has stats which were mostly distinct from Victreebel’s. Tsubomitto had an extra 15 points in Attack and Special Attack compared to Victreebel (120 vs 105 and 115 vs 100), but lost five points from Defense, Special Defense, and Speed, and 10 points from HP. Overall this made Tsubomitto a bit of a glass cannon, but given how small those stat drops are compared to the stat gains, Tsubomitto would have been significantly better than Victreebel with that distribution (aside: again, its stats don’t seem built on Victreebel’s, implying again that it was built from scratch). By July 30th, however, the team decided against this, and instead made Tsubomitto mostly the same as Victreebel, except that Tsubomitto had 15 more Attack and Victreebel had 15 more Special Attack. Given that—by this point—Tsubomitto’s only really unique difference with Victreebel was learning the Special Move Solar Beam at level 54, this feels like a pretty crippling downgrade. The back-and-forths the team went through with Tsubomitto’s moveset and stats highlight two things. First, it stresses that Tsubomitto was a very likely candidate for the final game. Given how much work the team put into tweaking its stats and its moveset, there was clearly a lot of attention and love for this guy. Beyond just the stats and moveset, Tsubomitto also got a redrawn sprite for June 1999. Even Steelix and Plux, other Pokemon in the same vein as Tsubomitto, more or less had the same sprite until the final (or when Plux was deleted), and yet Tsubomitto’s leaves required a touch-up. It even existed long enough to have a Pokedex entry! “It lives by absorbing nutrients in the ground through its root-like feet.” Okay it’s an incredibly boring Pokedex entry, but it’s cool that one exists at all! (Thanks to @RacieBeep for her take on the SW'97 Tsubomitto) Tsubomitto—alongside Plux—is more or less my candidate for “Most likely 'mon to make it to the final.” Like Plux, Tsubomitto survived it all the way to Spaceworld ’99, and if you just looked at stats, footprints, typing, name, etc, you wouldn’t be able to tell that Tsubomitto was going to be cut. The only clue is that Tsubomitto’s SW99 palette was now blue and yellow and its front sprite (but not its back sprite) was replaced with a very early version of Lanturn. This had to have been a very recent change, given how heavy Tsubomitto’s footprint still is in the game—Lanturn even still has its moveset!—which strikes me as about the last moment a Pokemon could feasibly have been cut. On the other hand, the second thing demonstrated by the back-and-forth Tsubomitto went through is how doomed it was from the start. As much as the team seemed devoted to Tsubomitto, it seems like they struggled with finding a way to make Tsubomitto unique from Victreebel. Unlike Bellossom, which was able to lose Oddish’s Poison-type, at no point in development was changing Tsubomitto’s type ever on the table. Instead, the team first tried to differentiate its moveset by giving Solar Beam and Lovely Kiss, then by making it into a glass cannon. Neither worked: maybe Lovely Kiss and Solar Beam were not deemed distinct enough, and the stats they had given Tsubomitto made it too much of an upgrade to Victreebel. Fixing either of these just made it more of the same, and by late July, it feels as though the team had given up: the Tsubomitto of July 30th technically learns Lovely Kiss, but only if it was an opponent’s ‘mon, and its stats are, at best, a minor tweak of Victreebel’s. Maybe an alternate evolution which was still pretty close to the original was good for one slot in Generation II’s Pokedex, but not for two slots. And in the end, the team liked Bellossom a little bit better. The final straw was Chinchou. Chinchou sprung up for the first time in Spaceworld ’99, filling in an empty slot of the Pokedex that had once been (briefly) held by Gurotesu but had been unused since at least June. Once the team came up with Chinchou, there was an incentive to give this new guy an evolution, to make him more interesting in the new game. Given that Tsubomitto wasn’t really a new idea (it just added onto a Gen I family) and that it was already in a troubled place, removing it and giving its slot to Lanturn was probably a no brainer. It's a shame, I love this guy. In another life, it was Bellossom that got the chopping block. In our universe, Tsubomitto never got a chance to throw a strike or make the critical pitch in the 9th inning. ID 438: ParaClosing out Period 3b is another baby Pokemon! Like most of the babies of Era III, Para was created to be an early evolution of a early-game, mid-popularity Pokemon from Gen I. Honestly, we’ve been through so much of these guys by this point, I don’t have a lot more to say about Para. But we’ll look over the phenomenon of Era III baby Pokemon one more time. Like I've said before, I think the babies of Era III were not chosen based on any criteria like popularity, or how well they'd fit into the gameplay; I think we're witnessing a brainstorm in which the team tried to make as many possible baby Pokemon to see what worked and what didn't. Because, if we're being honest, Paras is probably the last Pokemon on anyone's list that needs a new evolution. Look, I like Paras, but he’s probably not anyone’s favorite Pokemon, alright? I doubt Paras got much use in Generation I. You can find him rarely in Mt. Moon in Red and Green, at level eight; if you miss him there, you can find him in the Safari Zone at a more reasonable level. In Gen II, likewise, Paras can be caught pretty early, in the Ilex Forest, or in the bug catching contest. Since Paras is so rare in Mt. Moon, I doubt very many players bothered with it: after all, a weak dude with Leech Life is more or less outclassed by even Zubat at that point in the game. Likewise, maybe people caught it in the Safari Zone, but by that time, the work evolving it to Parasect feels a bit like a waste of your time. So I guess what I’m trying to say is that Para feels redundant given how niche it's evolved forms were. In the final game, I could see them given to the play in the Ilex Forest instead of Paras, but given how painful it is to use Paras in Generation II, I doubt this would have been a welcome change. Maybe the designers agreed with me, given that Para didn’t last past Spaceworld ’97. What’s slightly odd here is that most of the other baby Pokemon found in Era III—Monja, Pudi, Hinaazu—are derived from Pokemon much more solidly found in the middle of the game, at a high enough level that there could conceivably be a pre-evolution of these guys lurking around. Doduo and Growlithe, for instance, are found in their mid-twenties and could conceivably have had a earlier form at levels below that; Tangela is found in the mid-thirties. So Paras is a slight outlier in that it's caught at level eight, too young to conceivably have a pre-evolution. It kind of makes Para feel like an odd choice for a baby. I doubt the designers were thinking this far ahead about Para and its potential usefulness in a playthrough, at least not initially. Like I said above, I get the sense that all of the baby Pokemon in Era III were the result of a brainstorming session in which the team tried to come up with a design for babies of any Gen I family that could conceivably have one: they’d decide later which ones, if any, fit into the larger game. That’s probably why so few from Era III survive into the 1999 reboot of development. And I don’t think Para had enough of an interesting hook to support its inclusion in the final. There’s not a lot else that’s interesting about Para in Spaceworld ’97. It’s moveset is identical to Paras, just learned at a lower level. It lacks the Grass-type, implying that Paras only gets the type when it evolves; this is a little strange, given how Para is almost entirely a mushroom. You’d think, if anything, this little guy would be pure Grass-type before gaining the bug-type. On the other hand, @OrangeFrench suggests that the mushroom in Para's sprite isn't actually part of the creature: it's emerging from an egg left by the parent inside a mushroom. Thus, it's pure bug type because Para hasn't been infected yet. Probably the most interesting thing about Para is its sprite, and the lore that goes behind it. As I’m sure most of you know, Paras and Parasect have a particularly dark origin story. Based on a real-life fungus that takes over insects and turns them into zombies, Paras and Parasect are apparently insects that have a parasitic mushroom on their backs. According to the Pokedex, when Paras evolves into Parasect, it actually loses complete control of itself, becoming entirely moved by the mushroom on its back. A bit unnerving, right? It’s strange that Parasect is the series’ only zombie Pokemon (Insert here when someone delves up some obscure Gen VII Pokemon from an alternate timeline or whatever it is happens in those games UPDATE- It was Greavard I had forgotten! Thanks CainNKalos!). Except that story doesn’t really jive with Para, does it? If the idea of the lore is that the mushroom gets bigger and bigger until it envelops the Paras, then why is there a baby Para coming out of a huge mushroom? My personal guess is that these mushrooms are the spores that a Parasect leaves behind, and that the Paras actually grows out of the spores. But again, if that’s the case, then it isn’t really possible for the mushrooms to be a parasite on the back of Paras; instead, wouldn’t they be in a symbiotic relationship? The life cycle looks a lot more like this: Mushroom creates Para -> Para grows a mushroom on its back and becomes Paras -> the mushroom completely takes over -> the Parasect shoots out spores, growing more Paras It seems like a completely different understanding of the relationship between Paras and the mushrooms. On the other hand, if @OrangeFrench is right and the mushroom isn't actually part of Para, the life cycle looks more like this: Parent plants eggs inside a mushroom -> egg hatches, and the baby Para eats the mushroom for nourishment -> Para matures into Paras, but is infected by its initial food source -> Paras matures into Parasect. The parasite has fully taken over -> The zombie Parasect then lays eggs inside small mushrooms, repeating the cycle. Again, maybe they hadn’t thought this far, but were just sketching ideas for what a baby Paras looks like. This is all probably overthinking it. Interestingly, by the time of the final game, Paras' Gold Pokedex entry gives its own explanation about how Paras reproduces: "It is doused with mushroom spores when it is born. As its body grows, mushrooms sprout from its back." By the final, there's no more room for Para in the lifecycle: Paras is just covered in mushroom spores from birth. One last thing about the sprite. It’s not a big deal, but if you look closely at the mushrooms on the front and the back sprite, they look like they’re based on the same drawing. Both have the exact same angle, they share orange dots in the exact same places (though on the front side there’s some shading) and their outlines are colored in the exact same places. It’s pretty clear to me that the back sprite was just drawn on top of the front sprite, or vice versa, and then they were both touched up and the front was given a face. This could be an indication that there wasn’t a lot of work spent on these sprites, or it could just be part of the artist’s style, but it’s interesting enough to point out. UPDATE: On twitter, "The Ordinary Guy" alerted me to a second backsprite for Para, found on the Scratchpads for Togetic. Take a look! The new sprite, found on the left, is probably an updated backsprite, made later than the one in Spaceworld '97. I didn't notice it at first, but it's significantly improved: the hands are better and match the front sprite, there's better shading on the mushroom, and it isn't just a hastily edited version of the front sprite. For all those reasons, I'm relatively sure this backsprite was probably created after Spaceworld '97, not before. Which is interesting, because it suggests that Para had at least a little bit more work done before he was scrapped in the 1999 reboot. It also is more evidence of how the end of the Korean Index is noticeably unpolished. Anyway, Para was removed, eventually replaced by Togetic, which will be a story all of its own when we get there. I doubt there’s any relationship between the two: Togetic just happened to take over a slot that had previously housed Para. Otherwise, that’s all for this unused baby ‘mon. Next time, we start up Period 3c, and begin finishing up the Korean Index!
14 Comments
ID 432 Kingdra Late into the Korean Index, the Pokemon team decided to add Gold and Silver’s new Dragon-type Pokemon into the roster: Kingdra. Kingdra’s neat, and an obvious addition, given how few Dragon type Pokemon were in the first generation of games. While it’s too bad that Kingdra is locked behind a trade evolution in the final games—another example of how, for some reason, Game Freak really didn’t want people to experience their new Pokemon—Kingdra was a fine addition, and I’m pretty confident that from the moment it was designed, it was a lock for the final roster. My guess is that Kingdra was inspired by Steelix. Era III has demonstrated that it’s difficult to make an evolution of a Generation I Pokemon and still make it interesting enough to justify a spot in the Pokedex; Madame and Jaranra were probably removed because they were more or less superfluous, as they were just copies of their pre-evolutions with slightly higher stats. I suspect Blissey and Porygon2 were also on the chopping block for most of development for the same reason. On the other hand, Steelix did something really interesting with a Gen I evolution because the Steel type fundamentally changed how Onix worked. Plux and Scizor were later changed to match this concept, and I suspect Animon was created with a similar idea in mind. So Kingdra was an attempt for lightning to strike twice, and to develop another Gen I Pokemon in a completely new and interesting direction. I like Horsea and Seadra, but they’re pretty forgettable Pokemon in Generation I, because they kind of get lost in the myriad of Water Pokemon available in those games. Kingdra changes the game: now Horsea and Seadra hide a secret power only hinted at in their current forms. Let’s talk about the Dragon type for a second. Dragon was one of the last three types added to Generation I, alongside Bug and Ghost. In the internal data, Bug and Ghost appear at the very end of the used physical types and Dragon is at the very end of the special types. This heavily suggests they were the last added to the game. There are also a lot of other clues that suggest they were late additions. First, note that Bug has only four moves of its type in Generation I—String Shot, Pin Missile, Twinneedle, and Leech Life. All of them are very low power, suggesting a lack of time balancing or developing them, and Pin Missile and Twinneedle are both more or less signature moves for Beedrill, meaning that Leech Life is the only damage dealing bug attack learned by bugs that aren't Beedrill (Jolteon also learns Pin Missile, so it's not technically a signature move). None of this makes a lot of sense and strikes me as a rush to add a few bug moves at the last minutes of development. Likewise, Ghost only has three moves—Confuse Ray, Lick, and Nightshade—and only Lick does Ghost damage (Night Shade does set damage and ignores type resistance). Dragon, of course, has only one move—Dragon Rage—which does a set forty damage, meaning that Dragon-types have no way of taking advantage of Dragon’s type advantage in Generation I, a bizarre choice. In addition, Ghost was clearly created in a hurry: in the early data we have of the move list, all three of the Ghost moves were given the Ground type as a placeholder. On top of all of this, Ghost type was incorrectly programmed in Generation I so that it would be do zero damage to Psychic Pokemon, despite the anime expressing saying that the only way to beat Sabrina was to bring a Ghost type to beat her Psychic Pokemon. No one really noticed this mistake, because Lick was the only move affected by this mistake and it does almost no damage, so it was unlikely to be used against Psychic Pokemon. In Generation II they fixed this mistake and made it effective against Psychic, strongly suggesting this was how it was always meant to be. As an aside, I have my suspicions that both Ice and Rock were also created pretty late in development. These were both the moves right above Ghost/Bug and Dragon in the Physical and Special Internal type list, which suggests they were made before Ghost, Bug and Dragon but after the other types. But furthermore, notice that there are no pure Ice-types in Generation I—even Jynx is Ice/Psychic—and that every other Ice Pokemon is Water-type? Also, in the early data for the move pool, Aurora Beam was a Water-type move, and Haze and Mist were both Normal, not Ice like in the final. Without these, Ice would have only had two moves—Ice Beam and Blizzard—and those could easily have been Water earlier in development as well. Rock is even more suspect. It only has two moves in the final game: Rock Throw and Rock Slide. Neither of these were Rock-type in the early moveset data we have: Rock Slide was a Normal move (and possibly something else entirely, since its power and accuracy were also changed) while Rock Throw was a Ground move. Ground also only has four moves in the final, and all but one of which strangely do 100 base power: Earthquake (100 power), Bonemerang (50 power, hits twice), and Dig (100 power, takes two turns). If Rock Throw and (maybe) Rock Slide were originally Ground moves, it would give Ground much more diversity to its moves, but as it is, both types feel very underdeveloped. My guess is that very late in development, those two were split up for some reason. In splitting the Rock type from the Ground type (despite leaving most Rock types also with a secondary type of Ground), the team needed to switch around the typings of the gym leaders. It's worth noting that Giovanni was originally a flying type Gym leader (even his badge in the final looks like a feather!) and all the Pokemon in Brock's gym are Ground type. This suggests that before the Rock/Ground split, Brock was a Ground-type gym leader. Anyway, that’s a long tangent, but let’s get back to Dragon. If it was such a late addition, why was it added? My guess is that these last three types—Bug, Ghost, and Dragon—were added to fit certain gameplay and story purposes as the team was developing the games. Bugs were probably added as an early game, extra weak type, to give the player Pokemon to fight they generally had a type advantage to. Thus, the Vidirian forest could be a tutorial that could teach how type-effectiveness worked. They probably also united the early Viridian Forest encounters by giving them a singular type and teaching the player that certain Pokemon follow different patterns: Bug types, for instance, evolve very quickly! My guess is that Caterpie and Metapod were initially normal types, and Weedle, Kakuna, and Beedrill Poison, Poison, and Poison/Flying respectively. I have a suspicion (without proof) that Venomoth was initially a Flying/Psychic (Or Poison/Psychic) Pokemon, hence its odd inclusion in Sabrina's otherwise Psychically-inclined team. Ghost types were probably added because the Lavender Tower didn’t feel unique enough without the Ghost-type: it wouldn’t make sense that you can’t spot the Ghosts there without the Silph Scope if they were just boring Poison-types. And Dragon-type was probably created explicitly as a “boss” type to make the Lance battle even more intimidating—notice that Dragons are strong against all three starter types, which means the player needs to rely on new Pokemon they’ve caught along their journey. Lance’s two Dragonair wouldn’t be nearly as intimidating if they were water type and Venusaur could just rip through them. Of course, making Dragons super-rare boss mons used by the coolest trainer in the league made the typing super popular. But it was extremely underdeveloped, and the Pokemon team probably knew that they needed to develop it further when they decided on making a sequel. I’m a bit surprised they didn’t make even more Dragon Pokemon beyond Kingdra, though if the team thought of Dragon as a “boss” type than they may not have wanted to saturate the games with Dragons, to make them all the more intimidating when the player did face them. In the final games, the team also came up with the Larvitar family, which is strangely not a Dragon type despite being pseudo-legendary and only appearing in the end game. They also made Kingdra the signature Pokemon of Claire, which has the nice effect of distinguishing her from her cousin, Lance, who just uses Dragonite. The team also added two new Dragon type moves in Gold and Silver, allowing players to realize their power fantasies of…having a Dragon move that does damage? Honestly this feels like a no-brainer to me: Twister was added as a weak 40 power Dragon move that causes the opponent to Flinch (based on Japanese myth that dragons could fly in circles and generate tornadoes), and Dragonbreath was a stronger Dragon-move that could also paralyze. They also added Outrage, which was...just Thrash but Dragon type? It got better in later Generations. Interestingly, Twister was the signature of the Dratini line in Spaceworld ’97, while Dragonbreath was a signature of the Horsea line (hinting at their inner nature). In the final game, they took Dragonbreath from Kingdra and made it into a TM exclusive move, presumably because they wanted Claire to give the player something of value after wasting their time in the Dragon Den. I think it’d be much cooler if it belonged exclusively to Kingdra though. Finally, lets talk about Kingdra’s sprite, as well as some early flavor that got dropped. Kingdra, as you can see, originally looked a lot more like a Dragon than it does in the final game. In the final, Kingdra really doesn’t look dragon-ey. It’s been redesigned as a type of Seahorse known as a Sea Dragon, which is cute as an idea but doesn’t really translate into an imposing design. It’s a strange change, and it seems like a last minute one, since the Dragon-ey design lasted until Spaceworld ’99, just a couple months from the final. I think we can pretty much guess what happened here though. The early design of Kingdra is a classic Nishida sprite, and I’m positive she originally drew it. However, very few of Nishida’s early designs lasted into the final game without major or minor tweaking: See, for instance, the Hoppip line, Politoed, Marill, and maybe Sneasel. My guess is that by the time the team had gotten to Spaceworld ’99, her original sprite just didn’t quite fit the art style of the final game anymore, and she was asked to redraw it. You can tell that, at the very least, it needed to touched up at least a bit, since the old SW ’97 sprite didn’t really work with the new Gameboy Color style coloring (look at how weird its outline looks, for example). I bet Nishida was busy with other things, and so she let Sugimori redo the sprite. And since he was doing redrawing it, he decided to start from scratch and completely re-did the design. The final design is also pretty much exactly Sugimori’s art style, so my guess is not that they didn’t like Nishida’s dragoney sprite, just that if Sugimori was going to touch it up, he was going to try his hand at a new design rather than just work with someone else’s idea. One interesting note about this is that the scratchpads have an unfinished in-between version of Kingdra on them, which features a mix of the features of both Kingdra flavors. This could have been a sketch done by Sugimori right as he started to redo the design; maybe he wanted to see what his sea-horse like design looked like on Nishida's existing sprite. Below, FrenchOrange actually reconstructed this sprite, to show what it would look like finished: As a quick note, I also find the early Silver sprite pretty interesting. It’s not done by Nishida—its outline is completely black unlike the main sprite, as one indicator—but it’s clearly done by someone trying to match her original design. They also chose a pose for Kingdra very reminiscent of Seadra’s pose, making me think that this sprite was done by a more minor member of the team. Probably someone created the sprite to fit in with what else already existed. One interesting consequence of this late-in-the-day change is that Kingdra lost its ability to blow ink out of its nose. Though it’s hard to tell, in the early Nishida sprite, the mustache-like black lines may have supposed to have been ink! You can tell because it’s early Pokedex entries explicitly reference this part of its lore: Early Gold: It lives deep in the ocean. It makes itself unidentifiable by shrouding itself with self-made ink. Early Silver: It always covers its body with the ink it spits out, so no one has been able to confirm what it looks like. In addition to that, the early moveset for the Horsea family includes Octazooka, the signature move of Octillery. At first I thought that was strange, but I realized it was a gameplay mechanic meant to mirror how Kingdra can fire ink; after noticing that, it made a whole lot more sense! However, Sugimori’s version of Kingdra clearly lacks any ink-related capabilities, so the team had to change its Pokedex entries, and its move pool, to match. It’s a weird change, but one that seems to line up with everything else going on with Kingdra. I would have really preferred more Dragon Pokemon in Gold and Silver: Kingdra’s cool, but it certainly feels disappointing for new Dragons to be so rare and basically unusable in a normal playthrough. Here’s another place where I think they really could have dropped Magby or Elekid in favor of some other Dragon family to hunt down. ID 433: Riifi (Leafy, Leafee, Leafeon) The next three entries in the Korean Index are three new evolutions of Eevee, adding to the Flareon, Jolteon, and Vaporeon trio of Generation I. This is another very obvious idea; while it perfectly fits the larger sensibility of Era III to expand on existing Generation I Pokemon, I’m honestly surprised it took the team all the way until now to come up with this idea. If Slowpoke got an alternate evolution way back in Era I, surely it must have crossed the minds of the designers to give Eevee more evolutions are well? There’s a deceptive amount of changes that happened to these three Eevi (plural of Eevee) between Spaceworld ’97 and the final games. Not only did the design team cull this new trio from three to two, they also decided to completely change the type and moveset of one of the surviving two. In doing so, they went from a near triad of rock-paper-scissors types (Psychic is good against Poison, is good against Grass, but unfortunately Grass is not good against Psychic…) to a dyad of two types designed to be rivals (Psychic and Dark). So let’s first talk about Riifi, the one that got away. Riifi (which would almost certainly have been localized to Leafeon, so I’m going to use that name from now on) is the Grass-type Eevee evolution that we’re lacking from Red and Green. It’s got roots running from its legs and feet, it's cute, it's adorable; what else can you ask for from an Eevee evolution? Generation I didn’t include a Grass-type Eevee, but instead went with an Electric/Water/Fire triad of evolutions. Though this is at odds with the main triad of the games—Fire, Water, and Grass—as we’ve discussed before, Generation I was originally developed around the Electric/Water/Fire triad and only late in development was it switched to feature Grass instead; in fact, before we had a “Special Stat,” Pokemon just had a “Fire”, “Water”, and “Electric” stat, which maybe was their special defense to those three types, or maybe their proficiency with those types (or maybe both). It’s anyone’s guess why the team changed to focus on Grass instead of Electric (my guess is that it makes more sense for Fire to burn Grass than to burn…lightning?) but through most of Generation I’s development, the team had created triads that included Electric instead of Grass (and sometimes Ice instead of Water). Eevee was designed simply too early for it to have had a Grass evolution. But with the increased prominence of Grass, a Grass-type Eevee was all but obvious for the sequel games. From the sprite design, it looks like Atsuko Nishida designed this Eevee (Fujiwara designed the originals). She likely designed Umbreon and Espeon as well, which means these were probably all created together. I’m unsure, however, if they were ever considered to be a new trio, or if Nishida was just brainstorming new types for Eevee. It's possible that the latter was the case, because even though Leafeon was present for Spaceworld ’97, it already looks like the team was considering dropping it. Unlike Espeon and Umbreon, which were found right next to each other in the Pokedex, Leafeon is found at the very end in slot 251. The Pokemon surrounding Leafeon (Snubbull, Togepi, Aipom, and Sneasel) were all clearly shoved into the Pokedex at the last minute and hadn't yet been properly organized yet; unless Leafeon was supposed to be the Mew of the sequel games (unlikely, since Eevee can just use a Leaf Stone to evolve into it), that its Pokedex number is obviously a placeholder, and its placement suggests it was added to the Spaceworld '97 build at the very last second. This all seems to indicate that Leafeon, by Spaceworld ’97, was already the team’s least favorite of the three. And lo and behold, Leafeon doesn’t appear at all after Spaceworld ’97. Instead, in this slot is a placeholder sprite in June 1999 (interestingly, it still has a green palette; Celebi didn’t initially have a green palette, but Leafeon certainly did before it was removed from this slot). This placeholder was then, late in development, replaced by Celebi, which was designed by Yoshida, a completely new designer who was brought in to fill some unused slots while the rest of the team was busy with other tasks. There's probably no relationship between Leafeon and Celebi, other than that they shared the same slot. Why was Leafeon the least popular new Eevee? There’s probably no way for us to know, but we can speculate. We know that the Psychic type came out of Generation I as the most popular type: there are quotes from the development team that they designed Lugia to be a Psychic type because all the best Gen I Pokemon were Psychic, for instance. So that probably means that Espeon’s place in the new game was assured. My first guess is just that, if the team already liked Espeon, they may have felt like Grass didn't fit alongside Espeon--after all, Psychic feels like a rarer, more advanced typing, while Grass is common, like the original trio of Eeveelutions. Maybe this was why they eventually changed Umbreon from Poison-type to Dark. Or, maybe Nishida just wasn’t as happy with the design as she was with the other two. That’s easily as likely. Leafeon’s moveset is mostly what you’d expect from a Grass-type Eevee, but there are a few anomalies. First is that Leafeon’s got Morning Sun, that weird move that heals more if it’s used during the morning hours of the day. I talked about this in detail when we discussed Sunflora, but in Spaceworld ’97 the Chikorita line and Sunflora also have the move, while in the final, it’s basically unused: Espeon is the only Pokemon to have it, and Espeon learns is at a very high level. As I said before, it was likely replaced with Synthesis on most Pokemon because Synthesis interacted better with the move Sunny Day. I think it’s very likely that if Leafeon had made it to the final game, this move would have become Synthesis (not to mention that the Leafeon from Gen IV learns Synthesis at a similar level). It’s just a bit odd that Morning Sun, a move designed for Grass Pokemon like Leafeon, became the signature move of Espeon by the final. The other strange aspect to the moveset is that Leafeon learns Wrap at level 56, 7 levels before it learns its final move, Solarbeam. This is, frankly, bizarre. Wrap is a very weak move, made even weaker by changes in Generation II that prevented it stopping the other Pokemon from using moves. Wrap is one of the first moves a lot of Pokemon learn, not the second-to-last. The other Eeveelutions (in SW’97) all tend to learn an supporting move at this level, or a move that supports a secondary strategy for that ‘mon: Vaporeon learns Mist, Umbreon Smokescreen, Espeon Light Screen, Flareon Rage, and Jolteon Pin Missile. Wrap does fit vaguely into this pattern, but it’s much blander then these other moves. My speculation here is that the team threw a move into this slot without much thought, demonstrating how Leafeon was either added last minute or was getting less attention from the designers than the others. Leafeon also has placeholder stats, like most other Pokemon in Spaceworld ’97, but we can be pretty certain what its stats were supposed to be. Each of the Eeveelutions had the same stats, shuffled around. They are were defined by one stat they excelled at and had 130 Points in. Vaporeon had a high HP, Flareon had attack, and Jolteon Speed; in Generation II, by the final, Espeon has high Special Attack and Umbreon has high Special Defense. The only stat left after those, of course, is Defense. So by process of elimination, Leafeon would have had 130 Defense and middling states in everything else (except Special, which would have been 110 like the original Eevee trio). In fact, we know this to be the case, because in Generation IV, Leafeon got reincarnated! Behold: On the Cryptodex, I’ve had an ongoing debate about whether cut Generation II Pokemon inspired Generation IV Pokemon, and I’ve mostly said that I’m skeptical of any direct connection. Leafeon’s the only exception. They don’t look identical, but they're pretty close, and the main differences can probably be put to the differences between Generation II and IV’s art styles. In addition, Nishida probably designed both of them, and it seems impossible to me that Nishida wouldn’t at least have her old design in the back of her mind as she created Leafeon. Their movesets are vaguely similar as well, though since there’s so much distance in time between the two games, it’s hard to draw too many similarities. On the other hand, how many different ways can there be to make a Grass-type Eevee? It’s going to look like a cat/rabbit/fox, it’s going to be green, and it’ll probably have leaves or roots on its bodies. So it’s pretty inevitable that, as long as the team set out to make a Grass Eevee, it was going to look pretty similar. I’m not sure there’s much difference between “Nishida brought back her old design from Gen II and improved on it,” and “the team designed a new Leafeon and it just happened to share similarities with the cut Pokemon from Gen II.” But I am glad we eventually got a Grass-type Eevee. Now, if only we get the Bug-type and the Ghost-type Eevees, we’ll all be able to die happily. ID 434: Espeon Now it’s on to Espeon, which is my personal favorite Eeveelution (everyone has to have one, right?). Espeon’s a neat concept, and one of the most recognizable Pokemon from Generation II; I’m sure it’s probably one of the more popular versions of Eevee as well, due to its typing. Espeon went through the least changes of the three Gen II Eeveelutions, but there’s still a bunch to talk about. Let’s start! First of all, let’s talk about what’s up with it’s name. Espeon’s Japanese name is officially romanized as “Eifie” (I've used "Eefee" above) which is very close to the Japanese name for Eevee, “Iivui”. I’m not really sure what they were going for with that name. It’s English name is much more clever: Espeon is a play on ESP, or extrasensory perception, which sometimes gets used as “Esper” in Japanese (ESPer, or someone who uses extrasensory perception, or is just generally mystical). I really like the use of ESP as kind of a play on words. Bulbapedia has a lot of explanations for what Espeon’s based on, very few of which make any sense to me. It claims that Espeon is likely based on a Nekomata or a Bakeneko, two Japanese yokai that are essentially intelligent, trickster cat spirits. I don't see a ton in that connection, but Nekomata are supposed to be tricksters which grow a second tail after living a long time; that's clearly where Espeon's tails came from. A Nekomata and Bakeneko, respectively Bulbapedia also suggests a Sphinx cat, which is more likely in my eyes, since there’s some similarity there. The inspiration that seems most obvious to me is the Dungeons and Dragons monster, a Displacer Beast, which looks like a purple panther with tentacles and a long lithe figure. Final Fantasy redesigned these as "Coeurls" and it's likely the Game Freak team knew the monster from there. While Espeon doesn’t have the tentacles of a Displacer Beast, it shares psychic powers that allow it to attack the minds of its enemies. A Sphinx Cat and Espeon. Of course, there's one more explanation: the mythical creature Carbuncle from South American lore. While I doubt the team was too familiar with myths about Carbuncles, the creature was used both in Yugioh and Final Fantasy; it shares Espeon's general body type, ears, and the crystal on its head. Espeon’s sprite doesn't quite look like any of these, and is notably golden in its initial design. This was probably because early on in Pokemon's life, the color gold was associated with the Psychic type: Sabrina's gym badge in Japanese is the "Gold Badge," and both the Abra line and the Drowzee line were primarily Golden. Though Espeon mostly kept the same shape and design, it went through multiple rounds of refining. Its earliest form is really rough. The most obvious rough part is the tails: look at how awkwardly one of them is bent, and how it has a weird outline where the tail joins with Espeon’s rump. Honestly, looking at the sprite I’m a little bit suspicious that second tail was added in a revision, and there was an earlier Espeon sprite with only one tail; they probably added the tail to make it look more otherworldly, it to match the long whiskers on a Coeurl. The way that second tail is just sort of overlayed on top of the other and squeezed in to fit the border of the sprite feels like it could have been added separately. iAnother thing to note about this version of the sprite is Espeon’s two extra ears. Yes, those protuberances on its head are actually a third and fourth ear! An early Pokedex entry confirms it: “By using all four of its ears, it can sense everything going on around it even when its eyes are closed.” By the final version of that sprite, those extra ears have been turned into tufts of fur. That’s a little bit odd, since Espeon’s whole thing is its ability to sense extremely well, and having an extra set of ears would convey that idea pretty well. The team could have decided extra ears are kind of weird or creepy, but my guess is that the final sprite was touched up by someone other than the original artist, and they just misinterpreted the sprite. Poor Espeon lost her ears because someone wasn’t paying close enough attention! After that first SW’97 sprite, the sprite was given new shading to match the Gameboy Color, a new bluish-purple palette, and the sprite was touched up to fix the weirdness with its tails (though they still don’t look perfect). Finally, right before the game was finalized, the sprite was completely redrawn, though it kept its basic look: Espeon was now taller, you could see under her ears, and overall the sprite felt much more refined. If I had to guess, I think three different people worked on this sprite. The original Spaceworld ’97 sprite was probably Nishida’s—not only did she do the other two Eevees in SW’97, but in interviews she’s been credited with the creation of Espeon. However, I bet that someone else—probably one of the more minor spriters who didn’t create their own sprites but just touched up existing ones—added the new shading, palette, and tails to the 1999 sprite. Not only did the outlining style change—Nishida’s SW’97 sprite had yellow flecks in its outline while the new one was solid black—but the initial Silver sprite looks incredibly poorly done while also appearing to be an attempt at mimicking the Gold sprite’s proportions from a different angle. Nishida is a better artist than that sprite suggests, and the way the head looks so close to the Gold sprite but just rotated makes me think it was someone mimicking her style, not the real deal. The final sprite, on the other hand, seems to have been done by Sugimori: it has the proportions of a Sugimori sprite and in general has the high level of polish that his sprites tend to exhibit. It was probably the case that the team knew Nishida’s sprite was old and not up to the level of everything else in the final game, and it seems to be the case that Sugimori spent the last couple months updating a number of the older sprites that had come that far. Given how late in the day this final sprite was created, it’s another example of a sprite that was tweaked for Pokemon Crystal. Espeon didn’t undergo as radical of changes between Gold and Crystal as say, Sneasel, but it did gain a deeper purple palette, which has been its signature color ever since. In the realm of gameplay, there were a few subtle tweaks from Spaceworld ’97 that, overall, made Espeon into a much better Pokemon than it was originally conceived to be. For instance, originally Espeon was conceived of as the Eeveelution with High Special Defense. It’s original moveset matched this focus on defense, in that Espeon got to learn both Reflect and Light Screen and lacked strong attacking moves until high levels. Espeon didn’t learn Psybeam in the original Spaceworld ’97 moveset at all (it learns Psybeam at 36 in the final, a reasonable level that gives it a lot of use), and learned Psychic at level 63, much later than in the final, where it learns Psychic at 47. Consequently, Espeon’s best attacking move would have Confusion for most of the game, and it would have had relatively low Special Attack with which to use Confusion. Overall this would have made Espeon relatively weak. Also strange is that the original version of Espeon learned Bite; it wasn’t yet a Dark-type move during SW’97, but it’s pretty clear why the move was removed by the final. As mentioned above, in the final, Espeon gets Morning Sun as its final level-up move. In fact, Espeon’s the only Pokemon which learns Morning Sun, even though it really doesn’t fit Espeon all that well. My guess is that the team wanted Morning Sun used somewhere in the game, since they programmed it in, but because it didn’t really work with Sunny Day, they didn’t want it to be too common and confuse players. So they gave it to Espeon, made Espeon learn it so late that you’d probably not run into it in a normal playthrough, and called it a day. Update (Thanks to GoldS_TCRF for this added info!): In fact, Espeon didn't get Morning Sun in its moveset until very, very late in development: August 23rd, or six days after Spaceworld '99. It seems that one of the last things the team did was to give moves that were completely unused to various Pokemon, so that the moves weren't in the data for no reason. At the same time, Misdreavus got Pain Split--originally given to Jigglypuff, Chansey, and Norowara, though it really only fit Norowara--Pineco and Forretress got Spikes--Which was originally Murkrow's signature--and Sneasel got Beat Up--a move only put into the game very late in development. So my instinct that Morning Sun was just sort of thrown onto Espeon seems to have been correct. Espeon also originally had a different evolutionary method. Because friendship evolutions didn’t yet exist in Spaceworld ’97, the three new Eevi all evolved with Evolutionary stones, just like the initial three. Espeon used the mysterious Heart Stone, which was also used by Politoad and Natu (but doesn’t actually work in this demo); Leafeon used the Leaf Stone; and Umbreon used the Poison Stone (which also didn't work in the demo). Eventually, of course, the designers got rid of the Poison Stone and the Heart Stone, leaving just a Sun Stone as a new evolutionary stone. I assume that by that point in development, the designers wanted Espeon and Umbreon to be paralleled, so letting Espeon evolve via the Sun Stone was off the table (what stone would Umbreon use?). Instead, the team decided on letting Espeon evolve via High Friendship, and Umbreon via low friendship. We’ve talked about friendship evolutions before, when we discussed Blissey and Crobat in particular. The friendship mechanic was probably first invented for Pokemon Yellow, and given that the two games were being developed simultaneously, an idea that worked well for Yellow probably seemed like a good candidate to import to Gold and Silver. However, friendship in Yellow had an obvious purpose: since your Pikachu was the star of the show, friendship could be used to change how the player got along with the Pikachu. It was much less clear how friendship would have any effect in Gold and Silver, but it looks like the team settled on using it as a new evolution requirement. Friendship didn’t quite fit the pattern they’d done with other new evolutionary methods—all other new evolutions used a method that wasn’t available in Generation I to give the illusion that those evolved Pokemon always existed, you just didn’t have the right item. While friendship as a mechanic technically didn’t exist in Generation I, there was nothing stopping the player from doing all the things that would make an Eevee happier, like using it in battles and walking around with it in the team. But it was close enough, I’m sure, to fit the criteria for a new evolutionary method. It’s highly likely that Espeon and Umbreon were the first two Pokemon to experiment with this evolutionary method. The data we have shows that by April 1999, four Pokemon evolved via friendship: Blissey and Espeon by high friendship, Crobat and Umbreon by low friendship. Because of the parallel nature of Umbreon and Espeon, I think it’s more likely that this evolutionary method was conceived in the first place to create a mirror between the two of them, and then applied to Blissey because, well, its name is BLISSey, and Crobat, which gives off an angry vibe and needed a new evolutionary method. On top of this, when the two were given these friendship evolutions, their moveset was also changed so that Espeon's final ability was Return (higher damage with higher friendship) and Umbreon's was Frustration (higher damage with lower friendship). So for awhile, the team was really thinking of using Espeon and Umbreon to show off the friendship mechanic. (Thanks GoldS_TCRF!) Of course, like I mentioned in Crobat’s entry, a low friendship evolution is a terrible idea, gameplay-wise. In order to get minimum friendship, you’d have to purposefully knock out your own Pokemon again and again, which really goes against the spirit of Pokemon and the idea that these creatures are your friends. It would also be frustrating, because walking around with your Crobat or Umbreon would naturally raise its friendship, making it difficult to actually get its friendship to minimum numbers when you evolve it. There’s a brief period in development when both Umbreon and Espeon evolve with High Friendship (which probably wouldn’t have worked, given that you’d have to cancel the evolution your Eevee to ever let it evolve into the second Pokemon), but the team quickly figured out another, even more flavorful way to parallel the two Pokemon. In the final, of course, Umbreon evolves at night and Espeon in the day. Not only does that match Umbreon’s flavor incredibly well, but it also makes great use of the game’s day/night cycle, a mechanic the team was very excited to use. Like I said in the last entry, my guess is that Espeon was the most popular of the three new Eeveelution designs, and the only one that was certain to be included in the Pokedex. If that’s right, Espeon probably exerted inertia on the other two designs; once a Psychic Eevee was decided upon, any other Eevee’s in Gold and Silver would have to have some interaction with it. With that thought on the top of our minds, let’s move onto Umbreon. (Thanks to SeviYummy for the art) ID 435: Umbreon Umbreon’s also one of the coolest Eevee designs, and the second of the trio found in SW’97 to make it to the final game. However, unlike Espeon, which stayed more or less the same throughout development, Umbreon had a complete rethink of its entire concept! Let’s take a look. As I’ve hinted a bit in the last two entries (and I’m sure many of your know), Umbreon wasn’t initially a Dark-type Pokemon! Instead, it was a Poison-type. As weird as that seems in retrospect, it actually makes more sense as a trio of Eevee’s with Espeon and Leafeon. In the original trio, the three made an almost but not perfect rock-paper-scissors of weaknesses: Jolteon zapped Vaporeon which demolished Flareon, but Flareon was neutral against Jolteon. Likewise, in SW’97, Espeon could rock Umbreon’s world, Umbreon was great against Leafeon, but Leafeon didn’t do anything to Espeon. So at least initially, I can understand why the team decided on a Poison-type Eevee. Once you know that Umbreon was initially Poison type, a lot of things about it make more sense. First of all, consider its final Pokedex entry from Gold: “When agitated, this Pokémon protects itself by spraying poisonous sweat from its pores.” It doesn’t make a lot of sense that a Dark type would secrete Poison from its body; however, it makes perfect sense if Umbreon was Poison type! In addition, ever wonder about those rings on its body? They’re a direct reference to the blue-ringed Octopus, one of the most venomous creatures in the world! Because Umbreon was conceived as a Poison type, it’s move pool from Spaceworld ’97 was obviously almost completely different from its final move set: It’s interesting that Umbreon learned Acid Armor, since that was an extremely rare move in Generation I: only Vaporeon and the Grimers could learn it (and really, who out there used a Muk and managed to level it up to 60 when we were kids?). Otherwise, it learned the whole gamut of typical Poison-type moves, even Smokescreen, which was eventually given to Flareon. In the final, Umbreon’s moveset is of course designed to trap Pokemon with Mean Look or punish them for switching with Pursuit; it’s a Pokemon much more designed around debuffing the opponent and otherwise tanking them. Which is another difference from the final. Like I mentioned above with Espeon, originally Espeon and Umbreon had switched stats (not identical, mind you, but close). While Espeon was originally the Eevee specializing in Special Defense (making it, honestly, not very good), Umbreon was originally a frail, slow, Special Attacker. Which would also have been much worse, since in its Poison form all of its most powerful moves were Physical moves, and even in when it was changed to Dark-type its moves were, on the whole, weaker than the moves Espeon got. Switching their stats around was definitely for the best for both of them: Espeon became a speedy Special Sweeper, and while Umbreon was still slow, Defense was made its secondary stat, making it good at receiving hits. It also gives Umbreon an interesting identity. Most Dark-type Pokemon are either speedy or attackers; Umbreon fills a niche as a powerful tank that happens to use Dark moves. Part of me thinks the design team made this stat switch because they realized how useless Special Attack was on a Poison Pokemon. Unfortunately, the timeline doesn’t quite match up: Umbreon was already a Dark-type Pokemon before they swapped the stats of Espeon and Umbreon. Maybe the old stats originated from before we have data, back when Umbreon was still a Poison type, and they decided to change its stats at the same time they decided to make it Dark type, but got around to changing the typing first? It's possible, but less likely that they simply decided to change the stats well after Umbreon had become a Dark-type. UPDATE (Thanks GoldS_TCRF): It turns out that the team actually experimented with giving Umbreon back its Poison moves well after it became Dark-type in the 1999 reboot. By Spaceworld 1999 (in August), the team had replaced Frustration with Toxic (which would be quickly replaced with Moonlight by the final) and Thief with Acid (which would become Confuse Ray in the final). This didn't last long (they changed their minds by next week), but this moveset change suggests the team might have briefly had a change of heart. If they weren't sure whether Umbreon would be Dark or Poison type, they may have changed it to be the Special Defense Eevee just to leave their options open (again, because Special Attack would do nothing for Poison moves). Also possible: the team had come to the realization that in order to make the Psychic type fair, Psychic types should probably have weak defenses. Making Espeon’s defenses weaker necessitated it being the “Special Attack” Eevee, which in turn probably led to swapping Umbreon for sake of consistency. By making Umbreon Dark type, there was another strange consequence. I’ve mentioned before that Game Freak, whether to make their new types scarce on purpose or just because the Pokemon of those types were designed too late to balance, made basically all Steel and Dark types unobtainable before the post-game. As frustrating as that is, there was one single Dark type available before the Elite Four in Gold and Silver: Umbreon! It’s a bit annoying that to get Umbreon you have to make a choice between it and Espeon, but it’s really interesting that Umbreon was essentially the ambassador for the Dark type. For most players of Gold and Silver, Umbreon was the only Dark type they’d played with at all. Last thing I want to mention about Umbreon concerns its sprites. Interestingly, if you put Espeon and Umbreon’s sprites side by side, they follow a similar pattern. By June 1999, both were using a touched up version of their SW'97 sprite with a new palette; both also have similar body proportions. Take a look: Both sprites are definitely Nishida sprites, or at the very least drawn by the same person, given the similar proportions. Of course, Espeon got an updated front sprite in the final, while Gold's Umbreon sprite is still this same one. Seems like the team had time for Sugimori to go back and refine Espeon's sprite, but he didn't get a chance to do the same with Umbreon's, so they just settled on using the updated Spaceworld '97 sprite. Except it looks like Sugimori did redesign Umbreon's sprite for the final, just not the Gold sprite. The Silver sprite, on the other hand, has the same proportions and detail as Espeon's final Gold sprite, suggesting they were done by the same person: Compared to the previous sprites above, it's clear that one set was done by the same artist, and the other set was done by a different artist. My guess is that Sugimori focused on the Silver sprite for Umbreon because no one had made one yet, and they left the Gold sprite alone because he just didn't have time to do two new sprites for Umbreon. This is more or less confirmed by Crystal, which used the Silver sprite instead of the Gold sprite, signalling that the team thought the Silver sprite was the better (or more refined) of the two.
I know this is all really minor, but it once again drives home how close to the wire the team was in finalizing Gold and Silver. I also wonder if they kept the reference to blue-ringed octopi simply because they didn't have time to rethink Umbreon after they changed it to a Dark-type; that's certainly why its out-of-date Pokedex entry was used in the final. Umbreon wasn’t completely neglected, like some other designs that didn’t get changed until the last minute (Delibird, or Misdreavus, for instance). But who knows what the team could have done with it if they had had an extra month to touch it up? Still, Umbreon’s cool. One of my favorites. Maybe it would have been even cooler as a Poison type? What do you think? 5/10/2023 4 Comments ID 427 (Animon) to ID 431 (Pudi)ID 427: AnimonAnimon’s one of my favorite Pokemon found in the Spaceworld ’97 build. I love the goofy looking scream on its face, the cute yet fierce nature in its eyes, and I adore the idea of giving Ditto an evolution. With the spotty data we have from the Generation II leaks, Animon’s a bit of a mystery; saying that, I think we have enough evidence to be pretty confident in what Animon was supposed to do. (Art by @RacieBeep) We know that Animon was designed as an evolution for Ditto, because Spaceworld ’97 has evolution data which says that Ditto evolves when holding the Metal Coat and being traded. This is notably the same evolutionary method that Steelix, Scizor, and Plux were eventually given (all Pokemon that become Steel after evolving. Keep this on the backburner for now). Beyond that though, Animon is an odd Pokemon, because in all other ways it looks identical to Ditto. It’s Normal type, just like Ditto, and all it learns is Transform, just like Ditto. Its stats are all placeholders, so we don’t know if it would have gotten buffed in that regard, but raising Ditto’s stats is a weird endeavor. It doesn't really matter what Ditto's stats are, because it loses all of those stats the second it transforms. The whole puzzle of Animon raises a larger question: What could an evolution for Ditto even look like? Evolution, as a concept, is supposed to take a weaker Pokemon and make it stronger in some way, usually through stat improvements. But Ditto doesn’t grow stronger through gaining stats (except for HP). I’m sure the team at Game Freak were a bit puzzled by this problem as well. Ditto was a pretty iconic Pokemon from the first generation, and most of Era III was focused on expanding the lines of previously created Pokemon, so Ditto is an obvious candidate for an evolution. But while that strategy paid dividends with Steelix or Blissey, how is that formula supposed to work here? The key, I think, is that Ditto has a debilitating weakness, one which I’m sure the designers were aware of. Because Ditto needs to waste a turn to transform, that means that Ditto will always take one hit from its opponent before it can attack, essentially meaning that it will lose any mirror match. Given that Ditto’s whole gimmick was that it directly mirrored the power of its opponent, this was a pretty crippling Achilles' heel: if Ditto couldn’t even beat the Pokemon it transformed into, what even was the point of Ditto? Maybe an evolution for Ditto could solve this weakness? In Generation V, the designers solved this problem by given Ditto the ability “Imposter.” Imposter automatically transforms Ditto the turn it is swapped into a battle; that guarantees that Ditto will be able to fight its opponent on a level playing field. Unfortunately, Generation II didn’t have Pokemon Abilities, so a solution like this wouldn’t have worked. But maybe the team were thinking along these lines when they came up with Animon. There are a number of examples in Era III of Pokemon that aren’t quite finished. Obviously, whatever was planned for Taaban was never programmed into Spaceworld ’97. On top of that, all the Pokemon that evolve by evolutionary stone hadn’t yet had those stone evolutions programmed in (like Tsubomitto), and the Hitmon line’s evolution was still just a placeholder (remember that they all evolve from Tyrogue at the exact same time, and the only way to get a different Hitmon was to cancel the first evolution and wait for it to immediately evolve again). The reason I bring this up is to suggest that Animon was possibly also unfinished. While its evolutionary method is complete in Spaceworld '97, I think there was more going on here that the team just hadn’t yet had a chance to program. My theory is that the team was trying to come up with a way for Animon to transform into an improved version of the opponent's Pokemon. In particular, the way it evolves is suspicious: why use the Metal Coat, when that item was used for other Pokemon as a way to give them the Steel type after evolving? I think Animon was intended to replace the opponent’s type with the Steel type (Metal type, in this build) when it evolved. Maybe Animon was going to be Steel-type itself, and I’m not sure if it would have just replaced the primary type, the secondary type, or just add Steel type to whatever Pokemon it transformed to. But I’m relatively certain, given its design and its evolutionary method, that the intent was that Animon made metallic versions of the enemy Pokemon it copied. In this way, it would make better versions of the transformed Pokemon: because “Metal type” was envisioned as a primarily defensive type, it would usually mean that Animon was better equipped to handle the mirror match it created. It wouldn’t just replicate its opponents: it would make them stronger! Viewed with this theory in mind, Animon’s strange appearance begins to make more sense. It’s not just a weird looking Ditto: Animon’s a collection of iron filings, magnetized towards a metallic source. It’s pointy because that’s the direction that the iron filings are attracted to. The reason it gets magnetized like this is because, during the trade, Ditto fuses with the metal coat, becoming not only a blob but a metallic blob. Importantly, one thing to keep in mind is that in Spaceworld '97, Ditto is the only Pokemon to evolve with the Metal Coat; Onix still evolves by level. Which means that Metal Coat hadn't yet been fleshed out as an evolutionary method for Steel Pokemon. This weakens my case, admittedly, but it could have worked the other way around: even after Animon was cut from the games, the designers decided to use the one thing that worked from Animon and explicitly connect it to other Steel-type evolutions, like Steelix, Scizor, and Plux. Each of these Pokemon only gains the Steel type through a trade while holding the Metal Coat, and each of their designs are based on the idea that their previous evolution somehow became coated in metal during the trade process. The problem with this whole idea for Animon was, I’m sure, in the execution. How, exactly, would Animon keep its Steel-type when transforming? The easiest answer to this question was probably to create a “Transform+” sort of skill to give to Animon when it evolved, but then it’d have Transform as well (since Ditto learned it), which would feel awkward, and potentially create bad gameplay situations if the player accidentally used the wrong type of Transform. On top of that, Transform has always been infamously buggy; if you don’t believe me, try using Select to swap two moves after your Ditto transformed in Generation I. I’m sure there was some complication with the programming that involved swapping types during Transform. Maybe the team could have hard-coded Transform to preserve Animon’s type if it was Steel, but that seems like a worse solution that would have all the same problems as creating a new move to do it. If this really was the plan, my bet is that the team hadn’t decided on the best solution to this problem by the time Spaceworld ’97 was finished. On top of that, retaining the Steel-type came with its own problems. If Steel replaced the primary type of the copied Pokemon, Animon may not have the same-type bonus to the attacks it mimicked. It’d feel pretty bad for Animon to be worse at attacking than its counterpart. And even though Steel was a defensive-type, that didn’t mean it didn’t have weaknesses: it would really feel bad for Animon to transform into a Fire Pokemon, only to gain a weakness to the very fire moves its opponent used. These cases might have made the team unsure if Animon was even worth all the trouble to program: if the gameplay of retaining Steel-type didn’t feel good in the first place, what was the point? Amidst all of this, I think the team found a better solution that made the existence of Animon obsolete. Soon after the Spaceworld ’97 demo was created, Game Freak came up with a unique held item for Ditto: Metal Powder. Metal Powder, if held by a Ditto, increases its Defense and its Special Defense by 50%, even while transformed (only in Generation II). Effectively, this item solved the Ditto problem mentioned above in a simple and elegant way: yes, Ditto would always take one hit while transforming, but it’d have higher defensive stats than the opponent it copied, so it would still have a fighting chance against that opponent. Metal Powder didn't exist in Spaceworld '97; it only appears in the data after Animon was already removed, suggesting that it could have replaced Animon entirely. And it certainly did exactly what the team needed Animon to do; it even kept the flavor of a metallic Ditto intact! With this item in place, it’s no surprise that Animon disappeared from the roster by the time the team rebooted the game in 1999. Animon’s slot was replaced with Gligar, a unique Pokemon in its own right, and one that probably did a lot more for the diversity of Generation II than Animon ever could have. I get it. I love Animon, he’s one of my favorites from the Spaceworld ’97 demo. I would give a lot for him to have a place in the Pokemon roster. But I also understand why he was dropped. Animon’s a weirdo, and the problem he was tasked with solving was probably just done much better without the need for an entire evolution. ID 428: Porygon2 Speaking of weirdo Gen I Pokemon, the next Pokemon to get an evolution in Era III was Porygon! Porygon2 is the next generation of Porygon technology: smoother, more efficient, and ready for terraforming other planets (at least according to its final Pokedex entry; the draft is much less interesting and does not mention the colonization of other worlds). At least in its original form, Porygon2 didn't look anything like Porygon; it was a balloon animal lion. It was quite a strange choice. What was going on with this guy? (Art on left is by @RacieB; on the right is Sugimori's official art of Porygon2) Porygon was an odd duck in Generation I. The only way to get one was by paying an exorbitant amount of coins at the Game Corner. It was such a high amount that it could take the player hours of grinding to ever amass the money you needed to buy them (I know, because as a kid I fought the Elite Four over and over so I could eventually claim my Porygon prize). However, when you do get one, it turns out it’s actually pretty bad: Porygon has low stats, unimpressive typing, and though it learns Agility, Hyper Beam, and Recover, is usually overshadowed by Pokemon that are a lot easier to get. In fact, in an interview, Satoshi Tajiri mentioned that Porygon was meant to be nothing more than a throwaway joke: he was told over and over during development that the future was in 3D games, not in Gameboy games, so he created a 3D Pokemon of the future! It’s also probably the case that Porygon’s low stats were part of this joke. After spending all this time saving up for cutting edge technology, it turns out that the bird you caught in the wild was already better and more useful. (The peak of technological innovation!) Given how weak Porygon was, it was a prime candidate to get an evolution. And Porygon2 solves the problem of Porygon's weakness: by adding twenty to every stat, Porygon2 went from a measly 385 BST to an impressive 515 BST. Porygon2 has more or less the same movepool that Porygon did in Generation I (though it lost Hyper Beam). But on a stronger body, these moves hit much harder. If the theory of giving Porygon an evolution was sound, the practice was a little more complicated. The Pokemon team seemed to struggle to come up with a look for Porygon2, because their first attempt at it didn’t look anything at all like the original form. I remember when the Spaceworld ’97 leak first appeared, people didn’t even realize this was supposed to be Porygon2 at first. Instead, it just looks like a balloon lion. However, in retrospect, it’s quite clear what this design was going for. Porygon, of course, was made of angular Polygons. It was supposed to resemble the best that 3D graphics could do at the time, like how most N64 games look blocky and simplistic in their designs. Gaming engines at the time had a particular problem depicting circular objects. Because 3D objects were made by drawing vectors and filling them in to become a polygon, they were always going to be composed of flat sides. Since a sphere doesn’t have a flat side, the only way to make one would be to have an incredibly complicated polygon with an infinite number of incredibly tiny sides. Obviously, that would be difficult for any computer of the time to render. The fundamental concept for Porygon2 was to show the graphics of the future, or what Porygon would look like rendered on a much more powerful machine. What if it were made entirely out of spheres, something that the N64 of the time would have been unable to create? That, I think, is where the idea of this strange lion came from. If Porygon was made out of blocky polygons, Porygon2 was made entirely from spheres. Why a lion? I think the designer set out to create a Pokemon made of spheres first, and a lion became the most obvious animal that would look good made out of spheres. Porygon2’s mane, for instance, works well when made out of spheres, and its feline face works when created out of the same features. I doubt you’d get such an expressive face out of any other animal designs when you’re so limited in the shapes you can use. Speaking of the design, a lot of people have noticed a similarity between SW’97 Porygon2 and a pretty famous mascot for a company called Mr. Donuts in Japan. The mascot, Pon De Lion, admittedly does look almost exactly like Porygon2: However, as interesting as the resemblance is, it’s a total coincidence. Pon De Lion first debuted in 2003, six years after this design was discarded for Gold and Silver. There was an old rumor that the guy who created Panel De Lion, Yuichi Ito, may have worked on Gold and Silver during this period (and thus could have reused his design for Pon De Lion), but no one’s been able to confirm that rumor and its probably just not true. It’s a weird case of synchronicity, but there’s no relationship between the two designs. I think the design team realized that this design was a mistake, however, and by early 1999, they’d already decided to give Porygon2 a design that was much more reminiscent of Porygon. The team probably (correctly) surmised that Liongon was unrecognizable and threw away many of the charming parts of Porygon when it went in such a unique direction. The next version still had the same basic inspiration. Though it took a number of passes to perfect, the design was essentially to imagine what Porygon would look like on more advanced technology that was more adept at creating smooth surfaces. The revised Porygon2 isn’t entirely made of spheres, but it is made up of smooth surfaces, like spheres, that would be difficult for the technology of the time to render, showing off just how futuristic it was. It also heavily resembles the “drinking bird” toy, a popular toy that bobs its head in and out of water, that’s been parodied in a lot of popular culture. I have no idea if this was intentional. Honestly, I’m a little surprised that Porygon2 even made it past Spaceworld ’97. Remember that in 1998, the famous Porygon incident happened in Japan, giving dozens of kids seizures and nearly destroying the franchise before it had really even gotten off the ground. Because of that incident, to this day Porygon is persona non grata on the anime, being completely banned from any appearances. Given that this incident happened right in the middle of Gold and Silver’s hiatus (and maybe was even the cause of the hiatus and subsequent reboot), it seems natural that the team would have dropped Porygon’s evolution alongside the dozen other designs that didn’t make it into the 1999 reboot. It seems to me that Porygon2 must have had a champion on the design team that, say, Norowara did not. Porygon2 always evolved when given the Upgrade Disk item and traded. It isn’t that surprising for it to have a technology based evolution, but as an aside, can I quickly complain about just how many trade evolutions there are in Generation II? Besides Porygon2, Slowking, Steelix, Kingdra, and Scizor all needed to be traded to evolve (Plux and Animon would also have been on this list had they survived until the final game). There were four trade evolutions in Generation I, but at the very least those were four out of 150 new Pokemon to explore; here, some of the most interesting Gen II Pokemon, half the available Steel Pokemon, and the only new Dragon Pokemon were hidden behind an evolutionary method that would have been closed off for vast amounts of gamers. Combine this with how underused Slugma, Murkrow Misdreavus, Larvitar and Houndour were, it very much feels like the designers of Generation II were actively trying to keep the player from experiencing the new designs they’d created. Finally, let’s also address Porygon2’s name. Obviously, the name is a good fit for the Pokemon: Porygon2 feels exactly how you’d name the latest update for your computer. But what is also interesting is that the name follows exactly the format we know that the team used to as placeholders when designing Pokemon. For instance, the Sunfish Pokemon found in Spaceworld ’97 was named Manbo1, which essentially is the name of the species plus a number indicating the evolutionary level it was; the next Pokemon would presumably be named Manbo2 and Manbo3. In interviews, Sugimori has said that this naming scheme was common: before they had a name for the species, they’d often just name it [Type of animal][number in the evolutionary line]. Which, interestingly, exactly matches Porygon2’s name. This could mean nothing, but it could also mean that Porygon2 was originally just a placeholder name that the team decided fit well enough that they kept it. Or, similarly, made its name is a sort of in-joke among the team: by making it look exactly like they’d name placeholders in the index, it’s a joke that Porygon2 was designed as a direct upgrade to Porygon. Like I said, this could mean nothing; I still find it worth noting that the name itself might have some meaning behind it. I’m glad Porygon2 made it into Generation II; I’m not wild about either version of its design, but I’ve always loved Porygon and I’m glad this guy gets some love. He may be exiled from the anime for all time, but at least the game designers treat him with respect. (Wonderful Fan Art by 29Steph5) ID 429: Igglybuff And now…for another baby Pokemon! Igglybuff is probably the least interesting of the slew of baby Pokemon the team designed for SW’97. Staying mostly static throughout development, Igglybuff gives off a “paint by numbers” sort of vibe in its design: as if now that the team knew they needed more baby Pokemon, they just followed a template to create one for the Jigglypuff family. By this time in Era III, the team had already designed four baby Pokemon to fill out the roster: Betobebii, Doduno, Smoochum, and Magby. All four of those Pokemon were designed from relatively obscure or unpopular Generation I Pokemon, and Smoochum and Magby formed part of a trio of baby Pokemon that didn’t follow the patterns set by the rest (they evolved at a later, level, had higher stats, and learned most of the same moves their adult forms learned). Compared to them, Igglybuff is a reversion to the pattern set by Period 2a, in which the team had designed two babies as cuter versions of existing mascot Pokemon. It's even grouped together with them in the Spaceworld '97 Pokedex. Igglybuff, in particular, seems more or less a clone of Cleffa. They were more or less revised in concert with each other: Both of them got their first stats in April 1999, and they both follow a similar pattern in setting their stats (both lose from 5 to 25 stat points in all of their stats compared to their grown-up forms). They also both had their gender ratios adjusted at the same time, on August 30th, 1999, to be 25% male and 75% female. Both have all the same moves as their adult forms in Spaceworld ’97, like the rest of the babies. That's nothing notable, but when the game was rebooted in 1999, both of them lost all those moves except for a few weak beginning moves, and Sweet Kiss, which cannot be learned by their adult forms. Again, this is not notable (Pichu, for instance, follows the same moveset pattern). However, in the case of Cleffa and Igglybuff, changing their movesets in this way made them practically identical: One has Encore; the other has Defense Curl. Now, to be fair, Clefairy and Jigglypuff were always very similar Pokemon. But the way that these were changed at the same time, the similarity in movesets and stats, etc, just make me think that the designers created Igglybuff simply to make sure that the Clefairy and Jigglypuff families would continue to be counterparts. Thus, Igglybuff doesn’t seem like an inspired design; it feels like someone on the team checking off a box. This is likewise probably why, out of the ten babies designed in Era III, Igglybuff was one of only three that survived the 1999 reboot (the others being Smoochum and Magby, and we’ve talked about why they stayed on the roster). As long as Cleffa was going to be in the game, I suspect the team was committed to Jigglypuff having a baby form as well. Like Cleffa, Igglybuff’s sprites didn’t undergo a complete rework, like say Pichu; instead, they were slightly redone over the course of development, focusing most on its expression and the tuft on the top of its head. Both Cleffa and Igglybuff, interestingly, had their sprites revised almost entirely around these two aspects of their design; Cleffa also gained a nicer smile and the antenna on the top of its head changed significantly. I suspect Sugimori was the creator of both of these babies; this gradual revision of the sprites is very much his style; likewise, Sugimori was very likely to keep the original sprite and just modify parts on top of it as he revised, rather than creating a new sprite. In its initial form, Igglybuff was just a round ball with arms and legs (with a little spiral denoting the one curl of hair some babies have), but during the 1999 reboot, Igglybuff got a small tuft of fluff on its head. This is actually more significant than it at first appears, because if you focus on this tuft, you’ll notice that it looks like the end of a balloon that gets tied off after blowing it up. Jigglypuff had originally been designed to be a balloon Pokemon, but not much of that inspiration made it into Jigglypuff’s final sprite. I appreciate that Sugimori’s revisions were trying to bring Igglybuff visually closer to those origins. (Cleffa, likewise, underwent a similar revision. Clefairy’s final Gen I sprite was vaguely based on a star, but it doesn’t quite come through the sprite. Cleffa was lifewise revised to have more of a five-pointed star silhouette, to evoke that design inspiration more than Clefairy otherwise does.) Unfortunately, this nice nod to Igglybuff’s balloon inspirations was lost in the final sprite; Sugimori (or whoever revised it) presumably decided that the single tuft on the top of Igglybuff’s head was either aesthetically weird, or he decided that having the tied of end of the balloon was a bit too on the nose. The tuft was instead revised into a larger three pointed bob that more clearly looks like hair. In doing this, the spiral also lost its connotations as a baby's first curl of hair, since now Igglybuff had hair coming out of its head! While I do like the reference to balloons more, I have to admit that the final sprite looks just a bit better, so it was probably for the best. Igglybuff’s face also got revised during these tweaks, going from a strange and dopey looking smile, to an ecstatic open mouth, to a more controlled and arguably cuter smile for Crystal. Igglybuff’s cute, don’t get me wrong! But I do think that a lot of the other discarded babies would have been more interesting additions to Gold and Silver than Igglybuff. And again--like Cleffa--Igglybuff is great for making a cute Pokemon even cuter. But for gameplay, there isn’t really a reason for it to exist. ID 430: Koonya Koonya, on the other hand, is a bit more interesting than Igglybuff. Koonya has a number of odd features about it: it lasted longer than many of the other babies, it has some oddities to its sprite, and most interestingly, it is probably a remnant of Gen I’s design. Koonya’s cute! It’s one of the more significant losses from the SW’97 period of development. Koonya’s a baby version of Meowth. It’s an adorable kitten, so unlike Meowth, it’s sleeping rather than awake and alert like Meowth. It also has a coin on its head that’s worth a lot less than Meowth’s; TCRF tells me that Koonya’s coin is a 5-yen coin and Meowth’s is a Koban. It’s name is either a portmanteau of “ko” (child) and “nya” (meow), or it’s a combination of “Coin” and “Nya”. Beyond its pose, there’s not much difference between them. Koonya’s slightly smaller, but the overall body shape and characteristics are more or less the same. Given how different Meowth is to Persian, you'd expect a least a little differentiation, but the designers didn’t really go that route. Meowth was already pretty cute, and so maybe they weren’t sure how to make an even cuter version that still broadcast “cat” as well as Meowth did? Koonya's sprite does have one oddity: the coins floating over Koonya’s head make no sense, no matter how you look at them. Why are those coins there? They look a little like bits, or small laser shooting satellites from Japanese shoot ‘em ups, but why would a kitten have Bits (yes, Racie, pun intended)? It’s not like it’s a psychic type. Some people have suggested that the coins are just in the air above its head, but if that’s so, the sprite does a really bad job of depicting that. To add to that, Koonya’s backsprite, which otherwise matches the pose of its front sprite perfectly, doesn’t feature the coins at all, making it seem likely that the coins were somehow not supposed to be part of the final sprite. But if so, that’s strange too: no other sprites in Spaceworld ’97 have clearly unfinished parts like this. And notice that the sprite was touched up before June 1999, yet the coins were left in the composition; that makes it seem like they were supposed to be there. The most plausible explanation, as explained by TCRF, is that these coins were alternatives the artist drew because they weren’t sure which one would fit best on Koonya’s head. Which makes a lot of sense: The left coin is clearly the coin they’d already chosen, though the right coin looks so similar it’s almost indistinguishable. The middle coin, on the other hand, looks a bit more distinctive but also makes Koonya look more like Meowth, something they were probably smart to avoid. Also notice that the coin was slightly changed in the June 1999 touch up to the sprite, in that the middle dot was changed to a different color. It still doesn’t make sense why these coins are still in the June 1999 composition: maybe the person who touched it up was different from the original creator, and they didn’t know the coins weren’t supposed to be there? But given the available evidence, I’m liable to say this is most likely what happened. Koonya, at least according to common wisdom, is likely a holdover from Generation I’s development. In Red and Green's internal index, there is data for the backsprite of an unused Pokemon that looks like a baby Meowth: If this is true, it makes Koonya the last Pokemon in the Korean Index to date back to Generation I's development. Honestly, it feels out of place in Era III, which mostly avoided using old Pokemon designs. I have my suspicions that Madame was a holdover from Gen I, and I'll discuss my suspicions about Pudi in the next entry. Other than them though, everything in Era III seems designed specifically for Gold and Silver, which makes Koonya stand out. My best guess is that the team came to Koonya from the other end: they didn't design Koonya based on the unused ID 134, but they went looking for candidates for baby Pokemon and realized that Meowth would work well, leading them back to the idea of a Pre-Meowth design. Or maybe they just made a completely new Meowth baby form and it was coincidence that something similar existed in Generation I. After all, the poses of the unused ID 134 doesn't match Koonya's at all. It's tail is completely different from Koonya's, and it seems to have a different type of coin on its head than Koonya does. If Koonya was a Generation I import, they almost completely redesigned it, whereas Mikon, Gyopin, Puchikoon, and even Kotora all seem more or less identical, designwise, to the backsprites we have from Generation I. In fact, these drastic differences made me examine the Generation I backsprite even more, and I've started to wonder if it really was a Meowth Pre-evolution. ID 134 in the source code for Pokemon Blue does have a sprite for a Meowth-like Pokemon, but that Pokemon lacks any data that ties it definitively to Meowth. There’s no evolution data linking the two—though many of the unused Pokemon lack evolution data so this is not solid proof—and there’s no moveset data or cry that would link them together. Furthermore, we know that some of the unused backsprites from Gen I’s source code were probably not distinct Pokemon: ID 32 has a backsprite that’s almost certainly just an early backsprite for Nidoran Male or Nidorino, and ID 140’s backsprite is most likely an early unfinished backsprite for Magneton, not a distinct third form for the Magnemite family. The Pre-Meowth Sprite in ID 134 doesn’t look all that different from Meowth: what’s to say it wasn’t just an early Meowth backsprite that was left there when Meowth was shuffled earlier into the Internal Index? I mean, look at them side by side. If all we had were backsprites, is there anything that denotes the left one as a pre-evolution of the right one? Just going by back sprites, it seems far more likely that these backsprites were made for the same Pokemon, and the left one was just discarded. In fact, we know that Meowth's backsprite used to be larger. A lot of Pokemon that evolved later had their backsprites shrunken so they would look smaller in comparison to their evolved forms. Isn't a simpler explanation that this "Koonya" backsprite was just a Meowth backsprite before they shrunk it? There are some other oddities that make me question whether Koonya was ever in Generation I. First, let's pay attention to the stats of the Meowth line. I covered this in detail in the Puchikoon article, but in brief: In Generation I, Pokemon followed a strict pattern when evolving, usually gaining total stats in a set pattern. A second form Pokemon in a two-stage line--for instance, Koffing and Weezing--tend to get +125 to their total stats, while in a three-stage line, the second form got +75 and the third got +100 (sometimes +75 again). For examples of that second pattern, see the Nidoran families, Pidgeot, and the starter Pokemon. This pattern follows pretty consistently, but there are certain Pokemon lines that break the pattern. Some instances: the Zubat family, the Ponyta family, the Goldeen family, the Psyduck family, and a few more. In these two-stage families, the jump between the first form and the final form is not +125, like it would be in a two stage line, but +175. Importantly, if they were three stage families and one stage was missing, these stats would fit the general pattern. And, of course, that's what's so interesting: most of the Pokemon families that break this stat pattern were later discovered to have discarded evolutionary stages in Blue's source code! Their stats actually hint that one of their evolutions was cut in development (if only we had realized this before the leaks!). Except the Meowth line doesn’t follow this pattern. Persian gains 125 BST upon evolution, just like you’d expect a two-stage line to do. Which suggests that Meowth never had a pre-evolutionary form at all. Helix Chamber argues that this is just proof that Pre-Meowth was deleted sooner than the rest, or it was added and then deleted very late in the day, after stats were already allocated to Pokemon families. Both explanations are possible. But the simplest explanation is just that ID 134 was never a Meowth evolution, but just an early draft of Meowth itself. Furthermore, (and I know this is getting into the weeds and I apologize; please check out Helixchamber's article here for more specifics) a Pre-Meowth does not fit the patterns of Generation I's internal Index. ID 134 falls squarely in Gen I’s Period 4b, which was a series of Pokemon designs which were created as evolutionary relatives to Pokemon from Era I, the earliest creatures created (originally when the game was Capumon). But Meowth's not in this part of the Internal Index; its multiple slots later, in Era II (It's also weirdly misplaced in Era II, which makes it possible it was shuffled around from later in the Index). Saying that, if this was just an old slot for Meowth, it also wouldn't fit the established patterns: Persian was designed as part of Period 4b as well. None of what I'm saying is definitive, honestly. It's very possible, even probable, that ID 134 was a Meowth Pre-Evolution and not just a discarded Meowth backsprite. But given that Koonya was introduced in Era III of Gen II’s development rather than Period 2c-- when the team dug up the old Gen I designs for Puchikoon, Gyopin, and Mikon and made them into babies--it suggests more evidence that Koonya was a new Pokemon and not a holdover at all. I know we all want it to be true that there was a lost kitten in Generation I. And there very well may have been. But the evidence I’m seeing, while not conclusive, seems to suggest otherwise. The only other thing to note about Koonya is that it lasted a bit longer in the roster than most of the other baby Pokemon. When development was rebooted in 1999, ten baby Pokemon were culled from the list; it was clearly a priority for the team to move in a direction that featured much less baby Pokemon (probably because they’re useless in battle, obscure to find, and are too directly related to Generation I). The only ones that survived past this culling were all the baby Pokemon that made it to the final game, and funny enough, Koonya. Koonya’s sprite was even revised after the reboot to change its face and to give it a more pleasing palette. However, even though the sprite exists at least until June 1999, Koonya’s clearly already been dropped by that point. It’s typing is gone, as are any stats it may have been given; it doesn’t even have a name, and is instead found under “Mitei 05” (Pending 05). All this suggests that its sprite was being used as a placeholder until the team designed something new for that slot. The revised sprite suggests that there was a moment when Koonya was still going to be in the final games, but after June 1999, that doesn’t seem like a possibility. We know that Mitei 05 eventually became Pineco because Pineco’s shiny palette matches Koonya’s palette from June 1999. Pineco still had rough sprites even by the November 1999, when Spaceworld ’99 was built, also suggesting how late in the game the team came up with that idea. It was probably a good choice: Pineco could show off the new Headbutt mechanics found in the game, while Koonya…well, you might be able to breed one in the post-game. I get it, Koonya’s cute, but this one’s no real loss. More or less everything you get from Koonya existing, you already got from Meowth. I’m not in love with Pineco, but the diversity Pineco adds to the roster is by far a worthwhile trade for this cutie. ID 431: Pudi It’s Pudi! He’s a cute puppy! Designed to be a pre-evolution for Growlithe, Pudi lived the same life as many of the baby Pokemon from Spaceworld ’97 and was deleted without a trace during the 1999 reboot of Gold and Silver. According to TCRF, his name is probably a mix between puchi (petit) and Gaadi (Growlithe’s) name. Another baby Pokemon that didn’t quite make the cut, there’s two things worth discussing about Pudi. First, was Pudi one of the lost Gen I designs? And second, what exactly is the relationship between Pudi and the strange, cartoony doppleganger that looks similar to him, found at ID 325? This entry’s a little weird for me. Almost two years ago, I first published an article that covered ID 325, what I was calling Proto-Pudi, where I first dealt with the mystery connecting those sprites with Pudi’s design. That was two years ago! At the time, I didn’t dream that I’d ever be nearly at the end of the Korean Index, or that so many people would read what I had to say. So thanks everyone! I’m sure my analysis has improved considerably since then, but you might want to go back to that article to read what I had to say there about whether Pudi and Proto-Pudi are related. I still don’t think they are, at least not as directly as some people think. But let me elaborate my thoughts on this one more time. Proto-Pudi (if you don’t mind calling it that) is a weird sprite. It doesn’t look that much like a Pokemon at all, though it does have some pose and body similarities to Snubbull, and at least one reader has suggested that it might have had some relationship with Smeargle (just look at those noses right next to each other). What Proto-Pudi definitely is not, is an evolutionary relative to Growlithe and Arcanine. It doesn’t look anything like them: it sits on two legs while they’re four legged creatures, it has a horn, and its proportions are cartoony, not realistic like Growlithe and Arcanine. The only possible way they could be related is if Proto-Pudi were a very early experiment with baby Pokemon, and the first concept for baby Pokemon was to make them hyper-cartoony and unrealistic to accentuate their cute characteristics (even more so that Pichu or Cleffa, which still clearly fit, art style wise, next to the adult versions). At the same time, as I said two years ago, the similarities to Pudi are undeniable. They both have the same tail, the same floppy ears, and though Proto-Pudi’s mouth is much bigger and deformed, it is clearly the same type of mouth that Pudi has. There has to be some relationship between them. So here’s my theory. Proto-Pudi was designed as part of Period 1c, which was a transitory period. The Pokemon in 1c were usually more developed than the ones in 1a or 1b, though there were still a few oddballs that seem like they were only rough ideas. Most of the designs in 1c, if not all, were done by Atsuko Nishida. My guess, looking back on that part of the Korean Index, is that 1c was a space for Nishida’s early brainstorming sessions. She came up with all her initial ideas for the new game in this segment of the Korean Index, then decided which ones she would develop further and which she'd discard. Notice that almost all of the Pokemon in Period 1c got later redesigned, suggesting that even the ones that made it to the final were unrefined, rough ideas when they were first put into the Korean Index. Thus, Proto-Pudi was probably a very rough brainstorm, likely around the idea of a fun cartoony dog creature (a role that Snubbull eventually filled). My guess is that when Proto-Pudi was initially created, it had no connection to Growlithe or any other Pokemon. However, much later in Period 3b, all the designers were tasked to come up with evolutionary relatives for Generation I Pokemon, and in looking for inspiration, Nishida may have gone back through her early brainstorming sheet and found Proto-Pudi again. Realizing that the idea could work as a puppy, Nishida reworked the basic design into a pre-Growlithe, and then called it a day. Proto-Pudi had probably been abandoned a long time ago, and so from Nishida’s perspective, this unused design was a good place to draw from. This theory would account for their similarities while also explaining why Proto-Pudi just doesn’t fit in with the Growlithe line; I’m more or less content to believe this account of events as the most plausible reason these sprites look so similar yet so different. Proto-Pudi was related to Pudi, but through a much more round about way then we might have assumed. This question seemingly solved, let’s also discuss another, mutually exclusive, question about Pudi: was Pudi originally designed for Generation I? There are a few reasons to think this is also a possibility. First, it’d be a lot less strange that Koonya was pulled from Generation I if the very next Pokemon next to it in the Korean Index was also a holdover from Generation I. In fact, in Gen I’s internal index, there’s even a possible slot for where Pudi might have gone, right next to Pre-Meowth. If Koonya’s predecessor was in slot 134, 135 is notably blank; whatever was in there didn’t still have a back sprite in the Blue Source code. Since Pudi directly followed Koonya in the Korean Index, is it possible they were adapted to Generation II as a block, and Pudi was originally side by side with Pre-Meowth in Generation I? Furthermore, if Pudi did fill slot 135, it would continue to fit the pattern for Period 4b. Like I said above, Period 4b was mostly about creating evolutions for Pokemon designed in Period 1. And,like I said above, Pre-Meowth doesn’t quite fit this pattern (Meowth is found in Period 2!) but Pudi would, since Arcanine and Growlithe can both be found in the earliest stages of Generation I's development. And remember how Meowth and Persian’s stats don’t indicate that there was, at some point, a third form in their evolutionary family? Well, Growlithe and Arcanine do: Arcanine gains 175 BST when it evolves from Growlithe, a pattern that suggests at one point the Growlithe line evolved twice.
Weird, right? On top of that, the Growlithe line was the version exclusive counterpart for Vulpix, and in many other ways, these Pokemon were mirrored in Red and Green. It seems unlikely that the Vulpix family would ever have been considered as a three-stage line (if you include Mikon) but the Growlithe family didn't get the same consideration. Of course, they could have been made counterparts only retroactively—after all Vulpix seems to have been originally devised as part of a Water/Electric/Fire triad feature the Goldeen and Kotora families. But it's still suspicious. However, as convincing as all this evidence may be, Proto-Pudi seems to be the fly in the ointment to this theory. Because if Pudi really was a Generation I holdover, then why does it resemble Proto-Pudi, to the point that it seems designed as a revised version of Proto-Pudi? I don’t really have any good answer to this question; the existence of Proto-Pudi seems to definitively nix the idea that Pudi ever predated it. The only possible explanation that would make sense of this contradiction was if the Pudi from Generation I was so sketchy or undefined that it needed to be completely redone from scratch when it was imported into Generation II. Maybe it never even had sprites but was just planned: that could explain why Slot 135 has no backsprites in the Blue Source code. Alternatively, maybe Generation I’s pre-Growlithe looked much more similar to Proto-Pudi, for whatever reason, and Proto-Pudi is a more faithful rendering of what Pre-Growlithe looked like. I don’t know. The big problem with Pudi is that each of these explanations—that is, my explanation for the relationship between Pudi and Proto-Pudi, and my suggestion that Pudi was a Generation I design—convince me on their own of their veracity. However, they both seem unlikely to be true at the same time, as the evidence from either seems to contradict the other theory. Thus, it feels like only one or the other is true: either something completely different was going on with Proto-Pudi and Pudi was from Generation I, or Pudi was simply remade from Proto-Pudi and its design had nothing to do with Generation I. Whatever the truth is, it seems to still be a mystery. Anyway, you know the story with Pudi. Created for Spaceworld '97, Pudi was discarded with the majority of the rest of the baby Pokemon after the 1999 reboot. Pudi didn't really contribute a lot, it's cute but it's not very important for gameplay. You know the spiel by now. ID 422: MonjaMonja is the sadly lost and forgotten Tangela pre-evolution, found in Spaceworld ’97 but not elsewhere. Monja’s cute, a good extension of the themes of Tangela, and the only case of the team trying to make a single-stage Pokemon--Tangela--into a triple-evolutionary line. Whether related or not to this unique fact, Monja was also one of the very first to go when the team restarted development in early 1999: by the earliest data we have from that year, Monja was already completely erased from the roster. I’ve spoken before about the design philosophy of the Era III baby Pokemon. While earlier parts of the Korean Index sport baby Pokemon either designed from popular, cute Pokemon or repurposed unused Generation I as new baby pre-evolution, the babies in Era III tend to be designed around underused and outright unpopular Pokemon from Generation I. The purpose here might have been to find a way to make these Pokemon more interesting or usable; conversely, the purpose may have been simply to experiment with any Pokemon that conceivably had space in their evolutionary lines for a previously unseen family member. Tangela fits the profile of an "underused and unpopular" Pokemon to a T. Despite all the love Brian David Gilbert has recently thrown towards Tangela, Tangela is almost completely absent from Generation I. Barely any trainers use it, and the player can catch it only in one particularly obscure field of wild grass just south of Pallet Town, requiring Surf to even find. And even if you did catch Tangela, there’s no real purpose to using it. It starts with the weakest physical move in the game (Bind), it doesn't distinguish itself from the other grass Pokemon at all, and has middling stats. The only thing speaking to Tangela is that it’s the only pure-Grass Pokemon in that game, because for some reason almost every other one has Poison as a secondary type (I think this is a holdover from the development of the type system, but that’s a story for another day). (Credit to Chaoslindsey!) Giving Tangela a baby, then, makes sense: maybe by introducing Tangela as a more common Pokemon, earlier in the game, Tangela would find some more use, and in turn get more love. Going further than that, the team also decided to give it a third, evolved form: The fearsome Jaranra. By giving Tangela these two forms, Tangela could potentially be useful throughout the entirety of the game, making a footnote of a Pokemon from Generation I a potential key player in the sequel. Given how little of these babies and evolutions survived to the final game, it’s probably the case that the design team just had a change of heart about their design philosophy: rather than make Gold and Silver games which would further explore the Pokemon that already existed, the team instead decided to make more room for entirely new species. Given that change in direction, it’s no surprise that Monja didn’t make it to 1999, since Monja is, more or less, the Pokemon that most typifies the babies being created as part of this larger vision. We’ll discuss Jaranra in detail in the next entry, but it is worth pointing out here how Tangela is the only Pokemon which almost got promoted from a single stage line to a three stage evolutionary family. Later games, as they milked more and more possibilities for adding new evolutions of beloved ‘mons from earlier generations, have eventually caused a lot of older families to go from one stage to three stages: Chansey, Electabuzz, Magmar, and Mr. Mime are all examples straight off the top of my head. Tangela could have been the first to have been reinvented like this, and between Monja and Jaranra, I actually really like what the family could have looked like. It's not entirely clear why Tangela, out of all Pokemon, was the one chosen to get a baby Pokemon and an evolved form. For instance, why not choose Farfetch’d to get a baby Pokemon, to further fill out its family? Or Chansey? Given that the team was giving Chansey the evolved form of Blissey, why not also give it an earlier stage, like the team decided to do in Generation IV with Happi? I don’t really have an answer here, except that maybe the designer of Monja and Janranra had good ideas for what a baby and supersized Tangela would look like, and just didn’t have the same inspirations for others. Certainly, I can’t think of an obvious baby design for Farfetch’d, and the team clearly was having trouble coming up with a design that worked for Blissey. Of course, this might have been Monja and Jaranra’s undoing as well. Most of the babies—especially from Era III—didn’t make it into the final games, presumably because they took up a lot of Pokedex spaces that could have been used for entirely new, more interesting species. It would be bad enough that Madame was taking up a spot that could be used for Wooper; but these Tangela evolutions were taking up two slots in the Pokedex, for a Pokemon that wasn’t even that beloved in the previous games. If you’re looking for Pokemon to cut, Monja and Jaranra are easy first choices. Monja’s design is both reminiscent of Tangela’s, and interestingly different in some key ways. First to note is the clever wordplay of its name. Tangela’s Japanese name is Monjara; that means that Monja’s name follows the conventions they’d come up with for baby Pokemon by removing one syllable of the name. However, as TCRF points out, Monja’s name could also be a corruption of “mojamoja,” or unkempt hair (to be fair, Tangela's Monjara could also be a corruption of this). Monja also reflects the themes of Tangela, by showing a clever progression. Monja goes from messy hair to Tangela, which is completely covered in hair, to Jaranra, which is almost nothing but hair and a creepy smile. On the other hand, compared to Tangela, there’s some interesting differences. First of all, Monja’s a cyclops, while Tangela grows two eyes. Jaranra’s got none. This makes for an interesting play on the Sphinx’s riddle: What has one eye in the morning, two at midday, and none in the evening? It’s Tangela! I think there’s something more interesting about this eye, however. If you look at Tangela, the implication is that the eyes are hidden so far back in its hair that you can’t see what its face looks like (Diamond's Pokedex entry even makes this interpretation explicit: "It is shrouded by blue vines. No one has seen the face hidden behind this growth of vines.") Tangela could technically be anything under its hair; personally, I always imagined the Bill Clinton Tamagotchi: Monja, however, has a black, featureless body under its hair, suggesting that we actually had seen under Tangela’s hair, and that the black-featureless void of its face wasn’t because it was a mystery; that’s just how Tangela looks (maybe Monja's body is a seed, hence the visible body?). Which, personally, I think is a more boring interpretation of the sprite. The other main difference worth noticing are the sprites themselves: Obviously, Tangela stands out like a sore thumb. Not only is its palette completely different, but the art style also feels really distinct from the other two. There’s not a whole lot to say about this, except that this is probably because the sprites for the original 151 Pokemon were being handled completely differently from the brainstormed designs of the new additions. Monja and Jaranra were probably created by someone else besides Tangela’s original creator or the person who did the spritework for Tangela in Spaceworld ’97 (which, interestingly enough, was reused for the Japanese version of Pokemon Blue, whose updated front sprites were later used in the international versions of Pokemon). Monja and Jaranra have the same moveset as Tangela, but there is one unique change to Tangela’s moveset in SW'97 that I think is worth commenting on. In Spaceworld ’97, the Tangela family learns Night Shade (albeit at a very high level), a move that used to be the signature of the Ghastly family. It's, of course, a Ghost move, something very out of character for Tangela. In Spaceworld ’97, the team seemed to be experimenting with giving Night Shade to a more diverse set of Pokemon: in addition to the Tangela line, Spinarak and Ariados can learn it, and Norowara and Kyonpan can also learn it (Natu and Xatu get Night Shade in the final). Norowara makes sense because it’s the only additional Ghost Pokemon added to SW'97, but my sense is that the Tangelas and Spinarak’s got it because the team thought Night Shade could go to anything with a "creepy" aesthetic. And Jaranra, Tangela’s third form, certainly has “creepy face” vibes. Like I said above, Monja doesn’t appear after Spaceworld ’97; the evidence we have from the Scratchpads suggests that Sentret took its spot in the Pokedex, though it isn’t clear if Monja was dropped for Sentret, or if Sentret just took an already empty spot. We do know that Sentret was one of the earliest additions to the roster after the reboot of development in 1999. Not only does Sentret appear fully formed in the earliest 1999 data we have—it even has a full set of stats by April 1999—but Sentret also has its final palette already in place by June 1999 and its sprite is almost complete by that point as well. It’s evolution, Furret, is still using the palette of Gyopin, which it replaced, which suggests that Furret was a slightly later addition, but Sentret’s completeness suggests that Monja was one of the first SW ’97 Pokemon removed. And maybe Monja wasn’t that much of a loss; after all, a baby Tangela wouldn’t have added all that much to the gameplay of Generation II, and Tangela itself works well enough as an early game Pokemon. Monja looks cute, but I can understand why it was expendable. It’s nearest relative, on the other hand… (Thanks @Raciebeep!) ID 223: Jaranra Monja’s big sister, Jaranra, is a much more significant loss. Jaranra completes the Tangela family, replacing the dopey and kind Tangela with a monstrous adult. Jaranra, unfortunately, never made it into Generation II, but it left a legacy: Tangrowth, in Generation IV, seems to have some relationship to this lost design. (By the way, I have no idea what Jaranra's name might mean in Japanese, except that it's a further play on Monjara. If you have any idea, let me know!) Jaranra’s really cool. If Monja just has long hair, and Tangela’s is a mess, Jaranra’s design shows that she finally learned how to style the hair: no longer is it unkempt, but it now flows down Jaranra’s face like the “hime” (princess) hair cut that has been popular from time-to-time in Japan. Jaranra has bangs, flowing locks, and has its back hair done-up into a bow, reminiscent of some traditional haircuts associated with Geisha. I really like this detail; together, the hair on these three guys tells a story. (Notice the shortened front locks used to frame the face, and the longer hair behind it. Thanks to @Raciebeep for the art of Jaranra!) Jaranra’s hair, despite being styled, is still long; even longer than Tangela's. The hair now covers its body so much that instead of just two eyes peeking out, all you can see of Jaranra is its creepy, slightly hungry looking grin. It definitely is based on some sort of Yokai in Japanese folklore: to me, Jaranra looks most similar to a kuchisake-onna, a female spirit with a demonic grin. The Kuchisake-Onna has slits cut into its cheeks to make its smile extra large. If you meet one, it’ll ask you if it’s beautiful; if you answer no it’ll kill you, and if you answer yes, it’ll cut slits in your own face to make your smile resemble its own. Not only does Jaranra share the same creepy grin as this yokai, but traditionally, a kuchisake is depicted with more traditional Japanese hairstyles, like Jaranra's. Overall, it’s a pretty creepy origin story for Jaranra. (Credits: Left Isaac Huiza, Right Samice) If you want a slightly less nightmare-inducing origin, Jaranra’s smile also remind me of Noh Face from the Hayao Miyazaki movie Spirited Away. It’s definitely isn’t the inspiration, since Spirited Away came out a year after Gold and Silver were released, but if there was a traditional smiling yokai inspiration for Noh Face, than that might have been the shared origin for both it and Jaranra. Like Monja, Jaranra didn’t survive until the 1999 reboot of development. Unlike Monja, it left no trace at all in any of the files we have, which suggests that Jaranra might have been discarded even before it’s baby alter-ego. If Jaranra’s origins are the Kuchisake-Onna, than that might explain why it was so quickly discarded. First, we know that—due to the international popularity of Pokemon—the team was instructed halfway through development to make Pokemon more accessible for international audiences by downplaying particularly Japanese influences. That’s probably part of the reason Norowara, Kyonpan and Shigerufugu were discarded, and it probably played a role in how drastically Sneasel was redesigned. Secondly, there also seems to have been a push after 1999 to move away from darker, scarier designs and replace them with cuter Pokemon; Jaranra is not only kind of creepy herself, but she also has quite the scary backstory. Both of these could be the reason Jaranra was discarded, instead of the many final evolutions of Generation I Pokemon that accompanied it in Spaceworld ’97, such as Madame, Blissey, Tsubomitto, Belossom, Steelix, Scizor, and Plux. While all of Jaranra’s counterparts in Era III either made it into the final or were only discarded a month before the end of development, Jaranra is the odd one out by being such an early discarded design. Clearly, unlike the rest of Era III’s final evolutions, the team was most unhappy with Jaranra. These origins seem like the most likely reason. What really makes me despondent is that when Tangela did get an evolution, in Generation IV, the new design the team used was just so…awful. Jaranra, as I noted above, was a clever design that both showed development from Monja and Tangela and also added it’s own really unique twists on the design (most notably the smile). Tangrowth, is, well, a giant, fat, ugly version of Tangela. None of the subtleties of Jaranra’s design are present; whoever designed Tangrowth just took what we already had in Tangela and made it worse: I'm really just aghast that Tangrowth’s design is so bad while the team already had such a good design from years earlier. It’s likely a completely new designer took up the creation of Tangrowth; its unlikely that designer had access to the earlier design, or if they did, they clearly didn’t draw from it. It makes sense that a new artist would want to create something uniquely their own, rather than riff on a design some other artist made years ago, that wasn’t good enough even back when it was first created. I get that. They were probably told to create something that evoked nostalgia for Generation I, which Tangrowth, admittedly, does better than Jaranra. I just wish the new designer hadn’t created something so bad when there was such a homerun that escaped us. This discussion of Tangrowth connects to a larger conversation we need to have: the connection between Generation II and Generation IV. While Ruby and Sapphire were conceived of originally as a pseudo-reboot of the series—with an almost completely new roster and a bunch of Pokemon like Feebas that served the exact same in-game role as a Generation I counterpart—Generation IV was designed on purpose to appeal to the player’s nostalgia for Generations I and II. As a result, a lot of the new Pokemon added to Generation IV were evolutionary relatives of favorites from the first two generations. Mime Jr, Happiny, Munchlax, Magnezone, Rhyperior, Lickylicky, Electivire, Magmortar, Glaceon, Leafeon, and Porygon-Z were all new additions that evolved from Generation I Pokemon. For Generation II, the team introduced Ambipom, Mismagius, Weavile, Honchrow, Bonsly, Mantyke, Togekiss, Mammoswine and Yanmega. Since the moment Spaceworld ’97 was leaked, there has been speculation that the Generation IV throwbacks were actually designs reused from Generation II’s development. There are a number of parallels that make this convincing. First of all, Tangrowth and Lickilicky are evolutions for Tangela and Lickitung, the very same Pokemon that Jaranra and Nameru were designed to evolve from. On top of that, Leafeon closely resembles Riifi from Spaceworld ’97, and we know from other information (the internal filenames of Pokemon Battle Revolution) that Leafeon’s beta name was the exact same as Riifi’s. The Generation IV Pokemon that evolve from Gen II mons have less connection to anything in the Korean Index, but still, we have evidence that the Generation II team had briefly considered an extra evolution for Piloswine and Yanma (which Mammoswine and Yanmega now fill). In addition to this, there are a number of Generation IV Pokemon that have uncanny resemblances to some of the Korean Index designs that were dropped before Spaceworld ’97. As I’ve mentioned, there’s reason to suspect Chatot, Burmy, Mothim, and Cherrim might have been redesigned from some of the unused Korean Index mystery sprites. As fun as it is to speculate that Generation IV drew from all these designs, I have to say I’m skeptical. I don't think Lickilicky and Tangrowth have anything to do with Nameru and Jaranra, at the very least. Though it is absolutely possible these are just heavily redesigned versions of Nameru and Jaranra, there are a number of reasons to be suspicious. First, like I’ve said before, I have a hard time believing that the team would go backwards to failed designs so many years afterwards: while Generation III’s creation nearly overlapped with the development of Pokemon Crystal, Generation IV was released six years after Gold and Silver. How often do you go back to your creative projects six years later, especially ones you deemed not good enough the first time around? I also am deeply skeptical that the same designers who worked on Jaranra, Lickilicky, and Riifi worked on their Gen IV counterparts. Part of the reason is that their silouettes are completely different; in addition, the inspirations for each seem completely different. For instance, Tangrowth’s inspiration seems to be a bulky caveman (it and the other Pokemon that evolve upon learning Ancient Power take on prehistoric characteristics); that’s a completely different inspiration than the Kuchisake-Onna. While Jaranra has feminine features and clearly has learned to comb its hair, Tangrowth lacks any feminine features and its hair has just gotten more unruly. While I’d expect a Pokemon, redesigned after six years, to be almost completely different from its original appearance, I would also expect it to share the same design inspiration, just featured in very different ways. If they started from completely different origins, then I just don’t see any evidence one was iterated from the other. We’ll cover this as well when we get to Nameru and Riifi; for now, let’s just say that their original inspirations seem equally distant from the inspirations for their Gen IV counterparts. The design philosophy of Nameru and Jaranra is also completely different than the philosophy that animates the designs of Tangrowth and Lickilicky. In the case of Jaranra and Nameru, the designers seemed to want to find a new creature or idea that they could use as an inspiration to build on the features of the original Pokemon. For Tangela, a Kuchisake-Onna is a completely new concept unrelated to its Gen II origins. It works, however, as a base for Jaranra because it allowed the designers to style Jaranra’s hair in a way that was distinct (and showed growth) from Tangela’s. In the case of Nameru, Lickitung has absolutely no connections to a snake charmer, while Nameru was clearly modeled on the idea. The concept of a snake charmer, however, allowed the designers to play around with Lickitung’s most distinctive feature: it’s tongue, which now resembles a snake. But the design philosophy for Tangrowth and Lickilicky are entirely more boring. The designers seemingly wanted to capitalize on nostalgia for their first stages, by just making bigger, more exaggerated versions of the originals. Tangrowth doesn’t have much of a new concept (any swamp monster inspirations are very sublte); it’s just fundamentally a bigger version of Tangela. Likewise, Lickilicky takes the two characteristics of Lickitung—it likes to eat and has a long tongue—and exaggerates them, such that Lickilicky is now fatter than Lickitung and Lickilicky has a longer tongue. Jaranra and Nameru were designed to do something new; Lickilicky and Tangrowth were designed to exagerrate what was already present. Furthermore, if Generation IV did draw so heavily on the discarded designs of Generation II, why don’t we see more of them? Tsubomitto, Plux, and Madame were cut literally moments before the final build of Generation II, while Jaranra, Riifi, and Nameru were cut much earlier. Surely, if the designers were reaching back into their previous designs, they would’ve started with the designs that were almost good enough to be finalized, rather than revive the designs they had thrown out much earlier. I’d be more receptive to the connection between Generation II and IV if we saw a new Farfetch’d evolution, for instance, but as it is, the evidence seems thin. So why then do we see Tangrowth, Riifi, and Lickilicky in Generation IV at all? I think the answer’s pretty obvious: if the team was trying to extend some evolutionary lines from Generation I, Tangela and Lickitung are two of the most obvious candidates left from Gen I that still didn't have an evolution. After all, besides Farfetch’d and Weepinbell, how much more candidates really were there? They gave Rhydon, Electabuzz, Magmar, and Magneton all new evolutions as well; maybe Jynx, Persian, Dugtrio, and Lapras are all good candidates, but I have trouble believing any of those were as tempting as the one-stage and mostly-useless Lickitung and Tangela. As for Riifi: while I see some similarities, I think the idea of doing a Grass evolution of Eevee is more or less the most obvious Pokemon design idea in the world, and it doesn’t surprise me at all that both the Gen II and Gen IV designers independently came up with it. Yes, they do look similar, but how different do you really think two takes on “Grass-type Eevee” can look? So, in general, I think the similarities between Jaranra and Tangrowth (or Nameru and Lickilicky) are coincidences. Two different Pokemon design teams, separated by more than six years, happened to identify the same Generation I Pokemon as worthy of an evolutionary relative. At best, Sugimori or Nishida could have remembered that the team had tried and failed to create Tangela and Lickitung evolutions before; when they were brainstorming Pokemon to bring back for Generation IV, it’s very possible that one of the senior designers could have assigned the new developers to develop a new evolution based on a memory that they’d tried this idea before but weren’t happy with the result. I doubt those new designers even saw Jaranra and Nameru, but Sugimori or Nishida might be the connecting tissue. That’s not to say there’s no connection between Gen II and the returners from Gen IV. Like I’ve said before, I’m suspicious of a number of the Korean Index designs that didn’t get used at all: in particular, I’m almost positive that 382 and 383 are early versions of Burmy and Mothim, and I’m pretty suspicious about the Pokemon that may have become Cherrim as well. Likewise, it’s clear from development that the Gen II team wanted to give Piloswine and Yanma an evolution, but either didn’t have ideas for it or just simply didn’t have the space; I can very easily buy that they designed Mammoswine and Yanmega to fulfill the ambitions the earlier Gen II team was unable to meet for those two families, though the designs are probably dramatically different from how they would’ve looked in Generation II. And maybe Riifi was a straight redesign; we’ll discuss that when we get there. But I’m almost positive that the only connection between Jaranra and Tangrowth (or Nameru and Lickilicky) is that they were both attempts to give Tangela and Lickitung an extra form. Don’t get me wrong; Lickilicky and Tangrowth are so awful, I wish there was more of a connection between them and Nameru/Jaranra! Unfortunately, we’ll never see Jaranra, because Tangrowth took its place. Jaranra was obviously too interesting to make it into Pokemon canon. ID 424: Scizor Now let’s look at a final evolution design that DID make it to the final! At first glance, Scizor should’ve faced the same problems as Jaranra and Madame: it’s an evolution for a pretty obscure Generation I Pokemon, and one that doesn’t do much new or different. However, unlike those two, the team was able to modify Scizor into something interesting. By giving it the Steel type at the last minute, Scizor went from a tired design to a fascinating evolution almost overnight. Scizor is another "upgrade" design: An evolution for a single-stage Gen I Pokemon, probably meant to make that Pokemon more usable. Unlike Jaranra or Madame, Scizor had an advantage that made it more likely to survive until the final stage of development. While those two were evolutions of pretty obscure Pokemon, Scyther was a dark horse fan-favorite. Despite being also pretty obscure in Red and Green, Scyther had a pretty cool design (a vicious preying mantis with scythes for hands!) which made it popular. It also appeared prominently in an early episode of the anime—Dark City, which aired April 1998, right as development was paused. This probably put Scyther in a position where it had enough fans--both in Japan and in the design team--that when the team was putting the finishing touches on the roster, Scizor stayed, while its closest relatives on the list—Plux, Jaranra, Madame—didn’t quite make it across the finish line. To make Scizor work, however, the team made a number of drastic changes to its design late in the day that set it apart from Madame or Jaranra. Originally, Scizor was just a simply, straight upgrade to Scyther. It had identical typing, got 20 more Attack, 15 more speed, and 5 more special defense. Stat upgrades alone aren’t that interesting, and probably one of the reasons Madame never made it to the final; but even beyond that, these are very small upgrades to Scyther, making the original version of Scizor kind of a dud. The team may have known this, because by 1999, the development team changed everything about Scizor. Scizor gained a new palette, typing, and move set, making it a much different Pokemon than it started. The fundamental problem was that Scyther, as cool as it was, was a pretty bad Pokemon in Generation I. It had good stats, but Bug/Flying is a pretty poor typing, and it learned very few moves, including no STAB moves at all, even with TMs. My guess is that, given how cool Scyther was and how much more prominent it was—the second Gym leader, Bugsy, even had a memorable and difficult Scyther—the team was motivated to make Scizor strong enough that Scyther could overcome its fundamental weaknesses. So, possibly inspired by Steelix, the team decided to change Scizor’s typing to Bug/Steel. Instead of having five major weaknesses--as Scyther did--that small change made it so Scizor only had one weakness, to Fire. Scizor got slower—again, probably inspired by Steelix—but it gained attack and defense, overall a good trade for a much more defensively oriented Pokemon. The team then fixed the learnset by giving Scyther Wing Attack and Scizor Metal Claw, so now both of them had a STAB move. Finally, to bring it into line with Steelix—and because the design team wanted to make Steel extremely rare before beating the Elite Four—Scyther’s evolutionary requirements also changed. Instead of evolving after a certain level, Scyther now evolved the same way as Steelix, through being traded while carrying a Metal Coat. Basically, the team pasted what worked about Steelix onto a new Pokemon and called it a day. The sum total of this was to make Scizor way better and much easier to use. All of this was done late in develpoment: the change to Steel type was at the end of July 1999, right as the team had just given up on Madame and decided to give its slot up to Wooper. It’s clear that the team was struggling to make some of these final evolutions interesting enough to keep, and it’s likely that while they tried to find a reason to keep them, someone noticed how successful Steelix was and grafted the same idea onto Scizor. Steel was a new and cool typing; if the new evolution of Scyther was Steel-type, that was just enough of an innovation to make it interesting all on its own. Not to mention that Scizor got a complete overhaul in its looks! It became red instead of green like Scyther, and once it got the Steel-typing, there were changes to its sprite to make it look more metallic and robotic. Yes, the sprites stayed similar, and this was a touch-up more than anything. But with all the changes Scizor underwent just a few months before the final build, it’s fair to say Scizor was almost a complete rework. Comparing Scizor to Madame highlights just why Scizor was successful and Madame was not. Either could have been removed at the last second, but only Madame was, and I think the reason for that is because Madame just didn’t have enough of a hook to make it an exciting evolution to Farfetch’d. Even Blissey, another similar case, had a distinctive edge to Madame, in that Blissey had the maximum possible HP Stat and so was interesting in that regard. I suspect that had Madame stayed into the final game, the team would have found something to make it a more interesting evolution, whether a type change, a non-typical stat change, or something else. I would leave it there, but there’s one last weird oddity with Scizor’s development worth quickly commenting on. Notably, the early version of its Gold Pokedex entry is more or less identical to the final. But an early Silver Pokedex entry is just bizarre: “Since it prefers to spend its time by the waterside, it's conceivable that this Pokémon is related to Krabby.” What is heck is this? My first instinct was that this entry wasn’t meant for Scizor, but that it instead was written for another Pokemon, which Scizor overwrote. Maybe there was a crustacean Pokemon, previously unknown, that was briefly in a build between the one’s we have? Maybe it’s an early version of Corphish, which after all was made in Gen III and closely resembles Krabby? But that doesn’t make sense for a lot of reasons. First of all, Scizor didn’t overwrite any Pokemon; it had been in the game since at least the SW’97 build. So there was nothing else in that slot to have a Pokedex entry. Meaning that this entry was probably written while Scizor was in this Pokedex slot. I then thought that maybe Scizor switched Pokedex slots and overwrote something in its slot when they moved it, but that isn’t true either: it was Pokedex #233 since long before the team wrote Pokedex entries. Plus, the Pokedex entries were written pretty late in development, far past the time when Scizor might have been moved around. So the only likely conclusion is that this entry was meant for Scizor. It's weird, but it was probably a joke aimed at Scizor's design. After all, you can start to see reasons to think it fits Scizor. Scizor and Krabby are both red; they both have pincers; they both have an exoskeleton. So there are enough similarities, even though Scizor definitely doesn’t hang out on beaches and isn’t related to Krabby. I’m not sure we’ll ever really know what happened here, but my best guess is that this was a jokey entry written as a placeholder, before the team had thought up something to go into the Silver Pokedex. (Lost evolutionary relatives??) That's it for Scizor, a successful (and popular!) attempt to use a new evolution to put a cool twist on an exisitng Gen I design. Next, let's take a look at Purakkusu (Plux, or Plucks), and wonder why the same idea didn't quite work when used with Pinsir. ID 426: Purrakasu (Plux) And now let’s look at Scizor’s evil mirror universe twin. It’s Plux! The one that got away. We’re going to go out of order for a second here, largely because Plux (Purrakasu) has nearly the exact same development story as Scizor, and so it’s better to group them together. In Pokemon Red and Green, Pinsir and Scyther were version-exclusive counterparts: they were both beefy bug guys, both found in the Safari Zone and nowhere else, and they both had similar movesets. They were even named after a similar convention: Scyther was named after the Scythe, a cutting implement, and Pinsir after…pincers, another type of cutting implement. So when the team was developing extra evolutions for Gen I Pokemon, it made sense to give both these guys an upgrade. Maybe the designers did this to continue the parallels between the two Pokemon; maybe they created an evolution for both so they could decide which one they wanted to use once they had two concepts. Either way, as Scyther got a counterpart in Scizor, Pinsir got a counterpart in Plux. To continue the parallels, both were again named after “cutting” implements: Scizor is of course named after scissors, and Purakkusu was named after the act of plucking something with tweezers. Aesthetically, both were upgrades to the looks of the Gen I Pokemon. Scizor was more armored than Scyther, and replaced its scythes with scissor-like claws, which interestingly have eyes and look almost like they might be independently alive. Plux changed in altogether more interesting ways. It grew a third horn, for one, probably giving it more control over any victims that it grabbed in the pincers. Weirder, Plux gained a nondescript, white face, which looks a bit like a mask. It’s a bit off-putting, both because it looks so much like a blank stare, and because it gives Plux a very small, overtly human, face poking out from its very bug-like body. It terms of changes and tweaks during development, Plux is nearly the same story as Scizor. It had the same stat totals as Scizor throughout development, and it easily survived past Spaceworld ’97 into the 1999 period of development, alongside Scizor. Plux received a new palette in June 1999, just Scizor did, but Plux's palette looks a little bit wrong, probably in part because its sprite was still the one used in SW'97 and it hadn't been reshaded for a Gameboy Color style sprite (Scizor has the same palette problems, but they're less pronounced). Plux was also changed into Steel-type at the same time as Scizor, though unlike Scizor, Plux was never modified to learn a Steel-type move. Both Plux and Scizor also simultaneously changed their evolutionary method to evolve only when traded with a Metal Coat; presumably again because the developers wanted it to be difficult for the player to gain any Steel-type Pokemon before the post-game. They both, likewise, lost speed once they were redesigned as Steel-type Pokemon, with those speed stats reallocated to attack and defense. Up until Spaceworld ’99, both Plux and Scizor were shoe-ins into the final game. Plux even got not one, but two different Pokedex entries written for it, which heavily suggests the team considered it to be part of the final roster until very late. Here those entries are: “It pinches its opponents and flips them over. Its elongated horns contain steel components.” And “Taking advantage of its steel-like body, it lets itself take attacks, then pinches its opponent in its horns and flips them.” I’m not sure anything in those entries is revelatory (I think that’s how we all pictured Plux fighting), but it’s really cool that we have a little bit of lore for the guy. (Credit to @RacieBeep) Plux finally began to meet its end in Spaceworld ’99, where the Pokemon seems to have met a fate similar to Madame. Like Madame, Plux still had its name and typing in this build, and it still has the same backsprite and palette. However, Plux's frontsprite was replaced with a very sketchy first draft of Forretress (just like Madame's was replaced in the same build by a sketchy Wooper). From there, it looks as though Plux was replaced with the evolutionary relative of Pineco. Likely for the same reasons as Madame, the team decided this slot would be better used by adding an additional evolution to a new Generation II ‘mon than by giving a Gen I Pokemon and evolution. It's weird that the team decided that they didn’t want Plux, in the end; after all, it seemed locked in place until the very end of development, and Scizor made it into the final game just fine. At first, you might guess that maybe Plux was redundant: since it was so close in concept to Scizor, maybe the team decided they didn’t need two Bug/Steel evolutions. After all, this is probably what happened for them to discard Tsubomitto, which was a counterpart to Bellossom. However, if the team though they didn’t need a second Bug/Steel evolutionary ‘mon, they sure had a funny way of showing it. Because they replaced it with Forretress, a Bug/Steel evolution of another bug Pokemon. I’ve heard some arguments that Forretress was deemed an acceptable replacement because it filled a different niche than Plux: that while Plux was offensively oriented like Scizor, Forretress was more defensively oriented and so filled a more unique niche. And yes, Plux was a bit more offensively oriented. But the difference between the stats of Plux and Forretress are overstated. Though Forretress’s stats are lower than Plux’s, they are very similar in their spread: The only major difference is that Plux had higher attack than Forretress. But defensively, they’re more or less the same Pokemon, and both are equally slow. If we look at their level up moves, they have very different movesets, but the movesets amount to more or less the same thing. Neither have Bug or Steel type STAB attacks, both have moderately decent Normal moves. The main difference is that while Forretress explodes with Self-Destruct and Explosion, Plux gets Guillotine, Seismic Toss, and Swords Dance. So yes, their movesets do make Plux a different, more offensively-oriented type of Pokemon. But it’s easy to overstate that difference, as for the most part, Forretress has a lot in common with Plux and with a little tweaking, either of them could have filled the other’s role. There’s also the matter of Heracross. I wouldn’t be the first to note how similar Heracross and Plux are, both in aesthetics and in concept (fighting stag beetles). Upon first seeing Plux, many fans speculated that Heracross had simply replaced Plux and got transformed into a single-stage evolution. On the face of just the SW’97 evidence, this makes a lot of sense, and I would have bought the idea. Unfortunately, it’s not born out by the data we have from 1999. Interestingly, both Heracross and Plux coexisted in the same Pokedex, right next to each other! They were both also found ordered next to Scizor, implying that at one point, these three were considered to be a buff-bug trio. There’s an early sprite of Heracross that shows it with a much more bug-like face than its final sprite, which has led to some speculation that it was originally meant to evolve from Pinsir. However, that’s unlikely given that Plux was always in the Pokedex, even during Heracross’s early stages. Far more likely is that the team originally made Heracross more beetle like, but thought it could be more personable and likable with a dopey smiley face, like the final sprite. In general, that’s the big thing it has going for it that Plux does not: Plux is weird, kind of creepy, and alien while Heracross is a swell dude. Given that the 1999 phase of development was focused on making cuter and friendly designs, it seems obvious to me why they might have preferred Heracross over Plux. I do think Heracross likely made Plux redundant. Even though they both existed at the same time, the team had to realize that three bugs with similar move pools and similar aesthetics was probably too much, and Heracross is (arguably) the more interesting and superior design to Plux. Heracross also got to be Bug/Fighting, making it distinct from Scizor, and not simply a restatement of the same idea. Forretress might have been similar mechanically to Plux, but I wonder if the existence of Heracross pushed the team to create something that was aesthetically distinct from the other two bugs? We’ll never know for sure, but I think the loss of Plux is much more easily understood than say, Madame. However, both of these two were very close to making it into the final; unlike Jaranra, Plux and Madame were clearly worked on a lot before being discarded. If things had gone slightly differently, just think: we may have been discussing the strange way a popular Pokemon like Scyther never got an evolution even though the more obscure Plux became a Bug/Steel competitive Pokemon champion. ID 425: Magby After Smoochum was created, it was only a matter of time before Game Freak decided to complete the trio. Magby seems almost inevitable as a design: if Electabuzz and Jynx were going to get an early, baby version, then there was no way Magmar wasn’t. Much of what I’ve said about the other two applies to Magby as well, but let’s see what Magby adds to the conversation. As I’ve said before, Magby, Smoochum, and Elekid seem to be have been designed to be a different type of baby Pokemon than the others. Unlike Pichu and Cleffa, which are extra cute, learn much less moves, and are more or less helpless on their own, the Magby trio are all designed to look more like adolescents, all learn most of the same moves as their adult versions, have higher stat totals, and all evolve at a later level. Unlike the rest of the baby Pokemon, the Magby trio are conceived of more like a lost first evolution for Magmar and co. As a result, they function like independent Pokemon, unlike the rest of the babies. They’re weak, sure, but they still serve some battle purpose. Magby, for its part, looks like a smaller and cuter Magmar, just like you'd expect. In its initial SW'97 form, it had a kind of flame crown on its head. It looks good, but its incongruous with Magmar, which doesn't have anything comparable on its head. I can understand why it was removed. The other big difference in its early sprites was that Magby originally had a more segmented, armor-like, look to its hands and its arms, which seem closer to the design of Magmar than the final. It loses these in the final Gold sprite, but for some reason retained these hands in its back sprite and Silver sprite. This strikes me as a small tweak, done at the last second, and done sloppily enough that the team forgot to re-edit the other sprites. You can see how different the Magby trio is from the rest of the babies by comparing the original Elebebii design to the Elekid design that replaced it. Way back when, when we first discussed these two, I suggested that Elebebii/Elekid were not originally designed as a “baby” Pokemon, but part of the three-part evolution family. The big tip off for this is that they appear too early in the Korean Index, before the baby idea seems to have been invented; instead, they appear in a section where Sugimori was experimenting with different ways to give existing Pokemon a new evolutionary relative. The way Elebebii was discarded in favor of Elekid, and the way that it was later made into a baby Pokemon is, to some degree, a retcon to fit it into other ideas the designer had for Gold and Silver, created later in development. The main reason for the design team to prefer Elekid over Elebebii was probably because while Elebebii fit better as a normal baby Pokemon like Pichi and Cleffa, Elekid looks more independent and grown up, like the designs the team used for Smoochum and Magby. While during Spaceworld '97, the designers may have been using a design that fit better with the round balls Pichu and Cleffa, they may have switched to Elekid after realizing it could fit better with the other members of its triad. Elekid, of course, is not helpless: it sports hands, looks capable in a fight, and is much larger than Elebebii. Magby is certainly the cutest and most childlike of the trio, but it’s still much more grown up and adult than Elebebii was. The team might have made the decision to make this trio different around the time they came up with Smoochum; certainly, they had decided on this direction by the time Magby was created. I’ll never quite understand why this trio gets so much love and attention from Game Freak. Jynx is obviously a terrible design, but Electabuzz and Magmar are not really very good either. Both are ugly and strange looking, and pretty unpopular with Pokemon fandom in general. And yet, out of all the Pokemon in Gen I that could have had pre-evolutions, Game Freak decided that Smoochum, Magby, and Elekid were the ones they were dead set on keeping on the roster. Not Kotora; not Puchikoon, not Koonya, which would have certainly become a fan favorite. For some reason, Game Freak decided that all of these winning designs were superfluous and could be deleted; and yet somehow, wasting three slots in the Pokedex on Magby, Elekid, and Smoochum was essential. And then they did it again! In Generation IV, the team once again decided to create a third evolution for Magmar and Electabuzz. And these were even worse designs than the originals! Why waste valuable space in three different generations on these guys, especially when so many of the discarded Gen II designs were good? Part of this, I bet, can be put down to cultural differences. Magmar, Electabuzz, and Jynx all have roots in particular yokai from Japanese mythology. Magmar resembles as Basan, or fire-breathing Chicken “found” in Shikoku; it also seems to resemble a character found in Japanese theater stories called Hyottoko, which has a strange looking face (usually depicted by a mask) and is sometimes considered to be a God of fire. Electabuzz seems to be a depiction of a generic Japanese Oni, which is something of a cross between a demon and an Ogre. And as I discussed before, Jynx was probably designed to be a Yuki-Onna, or a female ghost that haunts people during the winter months. These sorts of references fall flat for international audiences, but I can only posit that the yokai origins of this trio make them interesting enough that the Pokemon designers have a fondness for them. And thus, taking up three whole slots of the new roster for their babies doesn’t seem quite so bizarre in a Japanese context. I’ve never seen data on the question, but I suspect that Magmar, Electabuzz and Jynx are far more popular in Japan than they are in the west. I also wonder if making these babies more independent, more useful, and more like real Pokemon was their saving grace. Betobebii, Puchikoon, Koonya, and the rest have zero use in battle. All of the babies, the Magby trio included, are extremely obscure--you can only get them from breeding--but at the very least Magby, Elekid, and Smoochum have a use once you do find them. When the design team was looking to cut designs, they may have noted how most of the babies were easily cut because they didn’t add anything to the gameplay of the games, but that Magby and co. at least, hypothetically, did. This one tiny difference in this trio might have been their saving grace. I can’t say I agree with this decision. Magby is the best designed of the three, in that it’s cute, endearing, and looks a bit like Magmar. It isn’t ugly and it doesn't knock you over the head with its theme like Elekid (did you realize Elekid conducts electricity?), nor is it giving the player the finger like Smoochum. But still, these are Pokemon you’d almost never run into in a normal playthrough, and even if you did, you wouldn’t want to. It seems to me that just about every scrapped design we’ve covered up to this point would have been a better fit in the final games than Magby, Elekid, or Smoochum. To me, this is probably the strangest decision the Pokemon team made in designing Gold and Silver. Period 3b OverviewPeriod 3b is the most coherent section of the Korean Index. Every design in this section is an evolutionary relative of a Generation I Pokemon (maybe excepting Taaban). It seems as though the development team thought the games didn’t have enough connections to the original games and used this section to create more. They did this by brainstorming new evolutionary forms for old Pokemon families. This could have been an idea that stemmed from the successes of Period 1d, which had featured multiple new evolutionary forms for Generation I (Politoed, Slowking, and Crobat in particular). These could have proven popular enough with the development team that decided to create more new and branching evolutions to existing ‘mons. Whatever the team's motivation, the next twenty Pokemon designs are tightly focused around this theme: none of them fall outside this pattern at all. We’re also finally done with unused designs: everything in the Korean Index from this point forward appeared in Spaceworld ’97. This is probably because this section was being completed only a few months earlier than the Spaceworld ’97 build, and all these last Pokemon were recent enough that they hadn't been overwritten but something else. Spaceworld ’97 was finished in November 1997; these designs were probably, at earliest, from April 1997 (six months earlier) and more likely even closer to the build date (potentially as close as two or three months earlier). That’s probably the reason many of their evolutionary methods were not yet programmed in in Spaceworld ’97: the team just had time to dump these Pokemon designs into the game and get them movesets before creating the SW’97 build. Things that would require more programming, like new evolutionary methods, were probably off the table at the time. ID 318: Taaban (Turban) ID 318, Taaban –whose name is clearly a Japanese transliteration of the word “Turban”—is one that we’ve talked about before, way back when we were discussing Slowking. I’m going to cover some of the same ground in this entry, but I encourage you to go back to that entry as well if you want to know more about this guy and read into some more detail about the theories I discuss here. Taaban’s quite a mystery. What purpose does it serve in the Spaceworld ’97 builds? It clearly looks like the Shellder attached to Slowbro and Slowking, but it doesn’t have any attributes connected to the Slowpoke family, and it doesn’t have any evolution data. It also has barely any excuse for a moveset, and nothing else for us to go by. Why did it get created? What purpose was it created for? What connection was it supposed to have with Slowbro and Slowking? First of all, Taaban was named after a Turban, but it was also named after the "Horned Turban" seashell, a type of shell that is conical jsut like Turban. This seashell has mythology around it of a horrible creature that lives inside the shell called a "Sazae Oni"; this creature looks a lot more frightening than Slowbro, but was probably the initial inspiration for the design. My guess is that Taaban was created, in part, to solve a continuity problem that had been created during the long, jumbled development of Generation I. One thing that just about every kid notices when they first evolve a Slowbro in Pokemon Red is that, even though the Pokedex calls the creature on its tail a “Shellder,” it looks nothing like a Shellder. Even an eight year old knows that a Shellder looks like a clam; whatever this thing that was on Slowbro’s tail was, it was conical and looked more like the shell that a hermit crab lives in than a clam. The anime did show the Shellder that grabbed a Slowpoke’s tail evolve into this form when the Slowpoke evolved with it, but the anime wasn’t fooling anyone. That wasn’t a Shellder on Slowbro’s tail, it was something else entirely. With the information we now know about Generation I’s development, we now know why this strange discontinuity happened in the first place. Shellder (originally looking like Cloyster) and Slowbro were some of the first monsters created for the original “Capumon” idea. Originally, these were mostly Kaiju like beasts on two legs that looked like something out of Ultraman or Godzilla; Slowbro was initially created as one of these Kaiju like beasts. In this form, having a shell on its back was just part of its design, and Shellder was a completely different, unrelated, Pokemon. Left: Shellder in the first Capumon Pitch Documents. Center: Shellder in Spaceworld '97. Right: Taaban in Spaceworld '97. It was only much later—When Slowbro was created the team didn’t even have a concept of evolution—that the team created Slowpoke as an earlier version of Slowbro. In creating Slowpoke, the team decided that the creature on Slowbro’s back was actually another creature, and that Slowpoke only evolved because it entered into a parasitic relationship with the clam on its back. And since Shellder was the only mollusk sort of Pokemon in Generation I, Game Freak decided to call Slowbro's mollusk friend a Shellder. And that's even though Shellder was created as a completely different creature and had a completely different design. So when Generation II began development, one of the designers might have been as concerned about this discontinuity as I was when I was a kid, and set out to create a solution. Taaban was probably made to retroactively explain why the Shellder on Slowbro (and now Slowking’s) tail looked so different from your regular average Shellder. It turns out, it was a different species altogether. Given that Taaban is found in Era III, it was probably also created as part of an initiative to create more Pokemon that evolve from or into Generation I Pokemon, to increase the links between the original games and their sequels. Which means that the design team was probably scanning through Generation I Pokemon that they could link to, found Slowbro, and realized they could kill two birds with one stone. Or even three birds, given that Slowking had already been created, and Taaban could be a link to him too. But how, exactly, would Taaban connect to these Pokemon? That seems to have been the question, and ultimately why Taaban was dropped. Notably, it’s not completely certain that Taaban was ever supposed to have a direct connection to the Slowpoke line, and given that nothing connecting it to them is programmed into SW ’97, the simplest possibility is that this wasn’t an accident. There are a lot of other possible scenarios where Taaban was just related to them through flavor, and not gameplay. The most obvious idea (which I somehow missed until a commenter pointed out this possibility) is that Taaban wasn't related to the Slowpokes at all, but was intended to be a branching evolution to Shellder. I buy this for a couple of reasons. First, a lot of stone evolutions didn't work in Spaceworld '97, and so if Taaban was meant to be a branching stone evolution, it might have been in the same boat as all of those and simply not programmed in. Secondly, this would explain Taaban's tiny moveset: it doesn't learn any moves because stone evolutions rarely learned moves in Generation I. In fact, Starmie, which is to some degree a close comparison, only learns Harden, Tackle, and Water Gun, all at level one. This is almost the exact same moveset that Taaban learns (Taaban gets Body Slam instead of Tackle)! There's definitely some evidence in favor of this idea. Honestly, I'm convincing myself writing about this idea, but there are two points that make me still doubt it. First, Taaban doesn't really look like a final-form Pokemon. Compared to Cloyster, which is much larger and more imposing than Shellder, Taaban is...about the same. It would feel very weird for Taaban to somehow be a upgrade to Shellder. Secondly, I can't get over how much Taaban resembles the shell the Slowpokes use. That feels too intentional. I've heard some other, more outlandish, ideas about how Taaban might have functioned separately from the Slowpoke lines. For instance, someone (I believe RacieBeep) suggested to me that Taaban might not have been intended to be directly related to Slowking, but instead an example of what happens when the Shellder on its tail was disconnected. She suggested that maybe Taaban was originally part of the Slowpoke tail subplot that appears in the final version of Gold and Silver. In that story, it’s revealed that Slowpoke tails are a local delicacy in Johto because they're so delicious; Team Rocket’s newest scheme turns out to be a plot to factory-farm Slowpokes and cut off their tails, so they can sell the tails on the open market. The Slowpoke tail is even an item, though all you can use it for is to sell it for a tidy sum; the story is already getting close to the uncomfortable question of whether people cook and eat Pokemon, and giving the tail an effect would have brought Slowpoke-cooking too close to the surface. RacieBeep (I believe) suggested that maybe the player could get a Taaban by somehow getting a Shellder to bite onto a severed Slowpoke Tail, which would then become its own independent Pokemon, free from the prison of a Slowking’s head (or a Slowbro’s butt, I guess). I could also imagine another possibilityt: maybe Taaban was just an early experiment in a regional form for Shellder, long before the idea of regional forms appeared in Gen VII. The Slowpoke Tail item appears in the SW ’97 build of Gold and Silver, so it is certainly plausible that Taaban was designed to go along with that sideplot. And it's clear that even in Spaceworld '97, Game Freak had wanted to spotlight Slowpoke in some way. There's no real evidence for this idea though, it's just fun speculation. Another idea suggested to me was that maybe Taaban was the result of a Slowbro's Shellder becoming unattached from its tail; the Slowbro would devolve back into a Slowpoke, but the Shellder might stay conical. However, given the other aspects of Taaban’s design, I think it’s far more likely that Taaban was designed to be involved in Slowking’s evolution, somehow. First of all, Taaban is clearly supposed to be the shell on Slowking’s head. It has the same ruby Slowking’s has on its backsprite and Slowking’s early backsprite has teeth from its shell driving straight into its head (making it look more independently alive). It also shares a general body plan and size with Slowbro's Shellder. To have it have no direct connection with the Slowpoke family would feel like one of the weirdest gaming teases in history! Secondly, even Taaban’s name suggests a connection with Slowking. Slowking is, after all, wearing its shell as a Turban, and so to name Taaban after its hat-like practically begs for players to make that connection. Thirdly, consider how sparse and uncreative Taaban’s moveset is. Taaban only learns three moves, all at level one: Harden, Body Slam, and Water Gun. None of these moves are particularly exciting. While this could be a placeholder moveset for an unfinished Pokemon, it seems unlikely: only Twinz and Girafarig had unfinished movesets in Spaceworld ’97, and otherwise the team had finalized movesets by the time of Spaceworld ’97. So this is probably the moveset the team wanted to go with. What this moveset suggests to me is that Taaban was designed as a gimmick Pokemon, like Unown, that didn’t serve a purpose in battle. The obvious gimmick is that it would assist Slowking (and maybe Slowbro) in evolving: the question is how? Taaban had nothing to do with the original intentions behind Slowking. As I mentioned in the Slowking article, Taaban hadn’t been created yet when Slowking was first designed, and Slowking was designed right around Politoed and Crobat, two other Pokemon created to be a new third-form evolution for a Generation I Pokemon. Slowking could have initially been designed as simple third form for Slowbro, in the same vein as Crobat, or it could have been designed to experiment with the idea of branching evolutions, like Politoed before it. Our evidence is contradictory on this point. On the one hand, early writing about Slowking from the writers of the Official Handbook (an early preview document released in 1996 which may have had information from the design team) indicated that he was supposed to evolve from Slowpoke: “Secret Data on Slowking: Slowpoke, a very stupid Pokémon, was out fishing for bait when a Shellder clamped onto on its tail, causing it to evolve into Slowbro. However, it is said that, in 1 instance out of 10,000, a Shellder will clamp down on a Slowpoke's head instead of its tail. As the Shellder bites down, its essence penetrates the Slowpoke's listless brain cells, bestowing upon it extreme motivation. Slowking has quickly become a hot topic among Pokémon collectors. How its evolution takes place is still unspecified, but it has been established that it evolves from Slowpoke!” On the other hand, in Spaceworld ’97, Slowking evolves from Slowbro, not Slowking. This could be an error, or it could be a temporary evolution method until the team figured out exactly how Slowking would branch from Slowpoke. Or, more simply, it could just be the way Slowking was intended to evolve at the time, before the team changed their mind and later made it into a branching evolution later in development. There are two likely possibilities concerning how Taaban fit into this scheme. First, it could simply be the case that the team hadn’t figured out what to do with Taaban yet. Taaban’s so late in the Korean Index that no Pokemon past this point are unused in Spaceworld ’97, so there’s the possibility that everything from this point forward was put into Spaceworld ’97 whether or not they had been completely figured out. Taaban could be an undercooked design that made it into SW’97 simply by being created so late. And if that was the case, the team may just have not yet thought through exactly how Taaban interacted with Slowking. Or it could even be the case that they weren’t even planning on any interaction and simply added in Taaban without planning to use it long term; Taaban could potentially be a placeholder. The second option is that Taaban had some as-yet-unprogrammed purpose in the evolution into Slowking. It’s possible that Slowking was programmed to evolve from Slowbro as a placeholder until they programmed in Taaban’s interaction with it. My guess, if this was what had happened, was that Taaban was involved in a trade evolution. Maybe, for instance, the designers planned that if a Slowpoke was traded to another player who was carrying a Taaban, both of them would merge together when the trade ended and give that player a Slowking. Another possibility is that Taaban was just supposed to facilitate an evolution of Slowpoke once it reached a certain level: say, at level twenty, if there was a Taaban in the party, then Slowpoke would start evolving (maybe Taaban would have survived the evolution, or maybe it would have been consumed as part of the process). Either way would work, though my bet’s on the trading mechanic. In the very next Generation of games, the team play around with this very concept: infamously, when Nincada evolves, it evolves into Ninjask and also leaves behind a Shedinja; Taaban might have been meant to be this idea but in reverse. Likewise, Mantyke was introduced in Generation IV, and functions almost exactly as I described above: it evolves into Mantine if there is a Remoraid in the party. So it's very clear that ideas such as these were floating around in the minds of the developers. If any of these ideas were the plan, however, it's very likely that they would have been difficult to implement. Not only would the programming to make these evolutions happen be complicated and unique, but there were probably a number of corner cases to work out. What happened to any item that Taaban was holding? Or maybe a player wouldn’t want to permanently lose their Taaban; there’s no easy way to indicate to the player that Taaban will be lost when the evolution starts, and it might feel bad for a player to allow the evolution to happen without knowing what will happen to their beloved shell-dude. Thirdly—and this is more a flavor problem—why does Taaban just evolve a Slowpoke into a Slowking and not Slowbro? After all, they have the same shell. They couldn’t retcon Slowbro’s evolution, since it was already established in the first game, so the reasoning for why you got a Slowking would have been confusing. It's likely that the messiness of these evolutionary methods was the reason Taaban was dropped after the Spaceworld ’97 build. Taaban’s slot was replaced with Piloswine in the next set of data we have, and Taaban was forgotten. There’s a slight possibility that Magcargo was loosely inspired by the idea of a living shell, but there’s no real evidence in favor of that idea. Taaban disappeared, regardless, and to this day Slowbro’s strange shell is still an odd continuity error. Art for this sad shell-dude by @Raciebeep! ID 319: Madame ID 319 is Madame, a masquerade mask wearing, Spring-Onion wielding, bank robbing, mysterious hero from the stars evolution to Farfetch’d. Madame is actually a fascinating Pokemon for two completely different reasons. First, we almost got Madame. Out of all of the unused Pokemon in Spaceworld ’97 Madame is part of a small group that not only survived into Spaceworld ’99, but also were very close to inclusion in the final game. Secondly, Madame is interesting because there’s a possibility it might in fact be a very old design. While we don’t know for sure, Madame might have been originally designed for Generation I, only to be culled from that game even earlier than the unused designs we have evidence of. Before delving into that second mystery, however, let’s discuss Madame’s history in Generation II. Madame is an evolution for Farfetch’d, and a cool looking one at that. She still has the signature green onion that Farfetch’d has, but it’s much larger, and Madame holds the onion like a baton, or maybe even like a samurai sword. She also now sports a mask over her eyes that brings to mind a masquerade ball. Its name, Madaamu or “Madame,” is also a connection to a masquerade, in that “Madame” is an upper-class way of talking about a lady that has heavy connotations of French Aristocracy. Essentially, Madame would be exactly the social class to attend a masquerade ball. I’m not sure how this would work with gender, since Farfetch’d is equally available in male and female forms, but that might have been something for the English localization team to work out as they came up with an English name. The sum total of this design is to present the audience with a classy, aristocratic art thief type character, in the vein of the "Gentlemen Thief" trope that you find in Japanese media like the anime Lupin III or the classy robber character that Catwoman exemplifies. You see this "mysterious gentlemen" character in a few other places. Most notably, Sailor Moon's Tuxedo Mask combines the mysterious, elegant nature of the character with a domino mask like Madame's; Speed Racer's Racer X or Mega Man's mysterious older brother Blues are the same sort of character, without the visual clues. Or, for a more recent example, Persona 5's main characters are all thieves, and the player character "Joker" and their partner Haru best resemble the "gentlemen thief" archetype that Madame was drawing on. Farfetch’d was originally designed around an old Japanese phrase “kamo ga negi o shottekuru” that translates to "a duck comes bearing spring onions.” The idea of the phrase is that this is incredibly lucky, since ducks are delicious when cooked with green onions; essentially, the duck drops in your lap the ingredients you need to cook it. The phrase means “easy target” or “lucky opportunity,” because of how fortuitous this is; it’s vaguely equivalent to the English phrase “shooting fish in a barrel.” Farfetch’d’s name in Japanese is actually just a contraction of this phrase, and it's suggested in the anime that Farfetch'd have been hunted close to extinction because of how delicious they are. It kind of feels like the Pokemon team went all in on this joke. In fact, I’ve seen people suggest before that in the final games, Farfetch’d is kind of a joke Pokemon in another way. Strangely, Farfetch’d is almost completely absent from Pokemon Red and Green. A few Bird Trainers use one while battling (strangely, a lot more trainers used them in earlier builds, like Team Rocket members and Gamblers), but there’s nowhere to catch a Farfetch’d; you can only trade a Spearow for a Farfetch’d named “DUX” in Vermillion City if you want one. But why would you? Its stats are terrible, and it’s not really an interesting or unique Pokemon, except for learning the overpowered move Sword’s Dance. The joke might be that you make a trade for a mysterious Pokemon someone is offering you in the game, and they give you a lemon; the joke's on you! Yellow version makes them catchable, but doesn’t change how unimpressive Farfetch’d is in battle. The thought process behind Madame probably sprung from this characteristic of Farfetch’d. It was okay that Farfetch’d was kind of awful in Red and Green, because it wasn’t a major part of that game, just a punchline more or less. But if the team wanted to feature Farfetch’d more in Gold and Silver, it would have to have more to it. And given that the final games have an entire section where the player chases after a mischievous Farfetch’d--and given that the team made Farfetch’d much more available for catching in Gold and Silver—Farfetch’d was already much more prominent. So Madame was probably an attempt to make Farfetch’d more useful, by giving it an evolution that could keep up in the later parts of the game. And Madame does exactly that: is has 40 more base attack, 20 more base HP and Defense, and more minor upgrades to Special Attack, Defense and Speed. Chasing a Farfetch'd Madame was also a good opportunity to create a link to Generation I, like the rest of Era III was trying to do, because the single-stage, low stat Farfetch’d was such an obvious candidate for an evolution. It also fit into Gold and Silver in basically the same way that Crobat did: both of them were an evolution that gave needed stats to an underpowered Generation I Pokemon, but otherwise didn’t add much to that Pokemon other than stats. Probably because it fit such an clean niche in the design, Madame was a mainstay in the Gold and Silver roster up until the very last minute. In fact, data connected to Madame still existed in the game until just about a month before the game was released! It's true! In fact, Madame not only survived past Spaceworld ’97—many Pokemon designs, of course, did not—but the designers touched up and redrew parts of its sprite, shaded it differently, gave it egg moves, TM moves, stats to go along with it, and even gave Madame a Pokedex entry. It’s a pretty boring Pokedex entry, mind you, and its vagueness suggests to me that it was still a placeholder, but it’s there: “A rare species that is seldom seen. Recent studies have discovered that the object it holds is a spring onion flower.” The team even changed the method that Farfetch’d evolved into Madame as late as July 21st, less than one month before Spaceworld ’99, indicating they still intended to use Madame that late into development. That July, they decided that rather than have Farfetch’d evolve at level 24, it would evolve with High Friendship, just like its counterpart Golbat. This was probably also a change to make sure that Farfetch’d evolved in a way that it couldn’t have in Generation I, to give the illusion that this evolution always existed in the previous games but the player couldn't activate it. This was a small detail that the team was very careful to preserve on all their new evolutions, and one that I've always very much liked. By Spaceworld ’99, however, something’s wrong. Madame is still Normal/Flying, still has all the same stats and moveset, but her frontsprites have been replaced by an extremely sketchy (and pretty goofy looking) first try at Wooper. The fact that all of Madame’s characteristics except her front sprites are still preserved suggests that this change must have happened pretty quickly, maybe just a few weeks before the build date of Spaceworld ’99 (August 1999). The creation of Wooper seems to have been a last minute decision, probably spawned from the need to have Quagsire or something with similar typing appear in low level areas. To give birth to Wooper, a sacrifice needed to be made. Not Madame In the last couple months of development, the team fully removed Madame to make room for Wooper, and Madame was cast into the dust bin of history. It was such a quick switch that Wooper even kept Madame's palette! Even stranger, in the final games, Wooper has a strange footprint, shaped like a duck foot. The team fixed this in later games, but now we know the truth: Wooper's footprint was one last final clue of the Madame that almost made it to the final games. For comparison, here's a picture of what Wooper's footprint looked like in Gold and Silver, and what it looked like in Crystal. It would take another twenty years for Farfetch’d to get the evolution he deserved, and even then, Sirfetch’d is a pale reflection of the majesty and grace of Madame. So what happened? Madame could have been removed because, at the end of the day, it’s pretty boring: Madame doesn’t add anything but stats to Farfetch’d, while something else in that slot could add a new Pokemon with new typing, or something that could diversify what the player could find in the wild. Admittedly, Crobat faces the same problem (being an evolution that only improves stats), as does Porygon 2 and a number of the other added evolutions from Era III. Not to mention that Wooper doesn’t add a lot more than Madame, given that it’s just a smaller Quagsire. Instead, I don’t think Madame was removed because there was anything inherently wrong with her; I think she was just removed to make way for Wooper. Quagsire is caught at a really low level in Spaceworld ’99, and he might have looked just a bit too imposing and chunky to be encountered at level 6, or fought at level 9 in a trainer battle. The team could have decided a smaller and cuter version of Quagsire made more sense, and so they looked for Pokemon to cut. Given that Farfetch’d was never the most popular Pokemon, and given that Wooper would be far more visible in the final game than an evolved Farfetch’d (after all, Madame would probably not appear in the wild or be wielded by trainers), it seems an easy decision to trade one for the other. This puts Madame in the company of two other Era III Pokemon that were deleted only at the very end of development: Tsubomitto and Plux. Tsubomitto, a split evolution of Weepinbell, was redesigned as late as June 1999 and still has some data surviving into Spaceworld ’99, but was ultimately removed by the final game and replaced with Lanturn. Plux was an evolution to Pinsir, meant to be a mirror to Scizor. It met the same fate as Madame: though its stats, name, and typing was intact by Spaceworld ’99, its frontsprites were replaced by goofy sketches of Forretress, which ultimately replaced it. Unlike Tsubomitto and Madame, it was replaced by a bug with its same typing, but otherwise, the pattern was clear: each of these three was replaced by an evolution of a new Pokemon from Generation II. The Almost-Were's This also suggests that the team may have felt that had gone too far in the direction of making Generation II an extension of Generation I, and that by the end of development the team wanted more diversity in their new designs. After all, by June 1999, the new roster was chock full of new evolutions for old Pokemon. Not only were Slowking, Politoed, and Crobat still part of the roster, but the roster still had Plux, Tsubomitto, and Madame alongside multiple other Era III designs like Blissey, Steelix, Kingdra, Scizor, Belossom, and Porygon2. That was probably just too many. Steelix and Kingdra probably had to stay because of their prominent place in the teams of two of the Gym leaders, but otherwise, I think all of these were on the chopping block to be replaced with evolutionary relatives of Gen II Pokemon. Tsubomitto and Plux probably got cut because their counterparts—Scizor evolves from Scyther, Bellossom from Gloom—had better designs, but Madame could’ve just been a random cut. We could just as easily be talking about whatever happened to Porygon2 or Blissey. There’s another important mystery pertaining to Madame, but it’ll take a bit of explaining. You’re probably familiar with the development of Pokemon Red and Green because you’re reading this (if you’re not, Helix Chamber does an amazing job summing this idea up here and here), but if you’re not, here’s a quick run down to get you up to speed. Essentially, there were originally 190 slots for Pokemon designs in Pokemon Red, which were filled haphazardly in the order that the Pokemon designs were created, much like the Korean Index. Basically, every time the designers came up with a new Pokemon, they programmed it into the next slot of the “Internal Index,” of Red and Green, which is still used in the final game to determine some of the features of any Pokemon the player catches. However, this “Internal Index” numbering system is not visible in the final game. Instead, the final games use the Pokedex ordering, which reordered Pokemon to be in a more logical order, putting evolutionary relatives together, legendary Pokemon at the end, and generally ordering Pokemon by how early in the game they’d be encountered. Of course, the final games only have 151 Pokemon. Before the games were finalized, this original Index was culled of around 39 designs that the team simply didn’t have space for. In their place, the final games have garbage data, which you can access if you use a variety of glitches; if you manage to access one of these unused Index slots, the game returns it as a “Missingno.”, a placeholder Pokemon whose stats and graphics are drawn from random data on the cartridge. For many years we just thought these unused Internal Index numbers were just a quirk of the programming of the game, but after the Gigaleak, we received backsprites and other data that proved that most of these were actually filled with discarded Pokemon designs at an earlier point of development. Anyway, Helix Chamber closely analyzed the patterns of the Internal Index and discovered that the Index revealed the logic behind the development of Pokemon Red and Green: you could see through the Index how the team came up with new ideas as development continued. Their work became the inspiration for what I’m doing in the Cryptodex, and in many ways what I’m doing here mirrors their method for Generation I. But interestingly, because their methods showed patterns in the Index, those patterns have helped suggest what these deleted Pokemon were like, and helped them guess what Pokemon would be in the deleted slots even before the leak revealed these backsprites. Again, you can read all of their analysis of the Internal Index here, but I've also abbreviated most of their findings in a poster I made that organized their analysis into sections (again, mirroring the sections of the Korean Index that I have delineated in the Cryptodex): Here's a zoomed in closeup of the two sections of the Internal Index that interest us today: Period 2b is from an era of Red and Green's development before evolution existed as a concept; it seems the designers had recently come up with the idea of different types and were experimenting with designs that exemplified the identities of these different types. For instance, as a poisonous land mine, Koffing represents what the Poison Type would look like, while Magneton demonstrates what a pure Electric type might look like. This section of the Internal Index was also concerned with creating regular animals that could populate the Pokemon world: Deer, Mankey, Diglett, Tauros, and Farfetch'd are all regular creatures given a Pokemon twist. Period 4a, on the other hand, is much later in development, after the designers had come up with the idea of evolution as a concept. As you can see, all of the Pokemon in this section are evolutions of Pokemon; in particular, they're all evolutions of Pokemon from Period 2b, which suggests that this section of the Index was about going back to designs created before the concept of evolution, and seeing if they could make those Pokemon into multi-stage families. This section is very tightly themed: everything in Period 4a relates directly back to Period 2b. But there's something strange in Period 4a. Notice the slots 115 and 121: those didn’t have a backsprite in the Gigaleak, which means that we had no data of what Pokemon used to fill those slots. They’re a mystery. They probably don’t have a backsprite in the leak because they were culled from the list earlier than the others; still, we have other evidence that they had, at one time, something in those slots. But the interesting thing about 115 and 121 is that, due to the fact that Period 4a has such a close relationship with Period 2b, we can make pretty educated conjectures about what they might have been. If everything in Period 4a of the Internal Index are evolutions of Pokémon found in Period 2b, 115 and 121 almost certainly were too. #115 and #121 thus have to be evolutions of Pokémon in that range. But what were they evolutions of? The candidates, interestingly, aren’t that numerous. There are 19 Pokemon in Period 2b which could have had an evolutionary relative in 115 or 121, and most of those we can cross off our list. Eight of the Pokemon in that range—Doduo, Koffing, baby Zubat, Mankey Diglett, Venonat, Seel, and the unknown Squid—already have evolutions in Period 4a, so we can discount them. That gives us eleven left. Poliwag has an evolution in Period 3b, so we're down to ten. As I’ve argued previously, Deer probably became Stantler in Gen II. There’s no discarded evolution for Stantler in the Korean Index, which means it’s unlikely that Deer originally had one (or else, why not also import it into the Korean Index?). Same with the #61 (The elephant that became ID 309 in the Korean Index), and Jagg (The shark which probably became #304, Ikari, in the Korean Index). And Magneton is unlikely because Magnemite appears later in the Index, and something related to the Magnet family is in slot 140; I doubt there was another Magneton evolution hiding elsewhere. So we’re down to six possibilities: Tauros, Crocky, Dragonite, Cactus, Farfetch’d and Jynx. I’m skeptical that Jynx or Dragonite ever had an evolution. Dragonite looks formidable as is, and later got two pre-evolutions, so I doubt it somehow became an even bigger dragon at one point. And though Jynx may not have been part of the Magmar/Electabuzz trio at that part of development, my guess is that association was already assumed by the time the team was development Period 4a. Thus, I think it's safe to say that 115 and 121 have to have been an evolution of one these last four candidates: Farfetch’d, Tauros, Crocky, and Cactus. And while those Internal Index slots could conceivably have held any of these Pokemon, a Farfetch’d evolution is by far the least...farfetched. In part, this is because I suspect Crocky and Cactus were on the chopping block for a long time, and in part because we found a design for a Farfetch'd evolution in the Korean Index, while we didn't find one for Tauros. This is by no means 100% certain. However I believe with a high probability that Madame, or something that was redesigned into Madame, held at least one of these slots. Helixchamber, independently and earlier than me, also came to the same conclusion and included Madame in their Gen I hack that included their interpretations of the unused Gen I mons. Helix Chamber's speculative Madame sprites, made by @catstorm (Orange) and @Raciebeep (Green) This is really cool, when you think about it! We may not have a backsprite, but we’ve solved the mystery of another of the slots of the Internal Index! My guess is that Madame didn’t make it into Generation I because of how peripheral Farfetch’d was in the games. If you think about it, Farfetch’d was probably surplus for requirements in Generation I: it was a Normal/Flying type in a game that also had Pidgey, Spearow, and Doduo. What more could Farfetch’d add? Farfetch’d probably just barely squeezed by the final cut of 151 Pokemon, and only because the team thought it was a funny joke: even then, it was included almost as a cameo. Since Farfetch’d was so unnecessary, Madame was probably deleted much earlier, and that’s why it doesn’t appear at all in the data we have. My guess is that the other Index number in Period 4a—whichever of 115 or 121 Madame didn’t occupy—was probably held by an evolution of Cactus or Crocky, and it got deleted for the same reason. The only difference was that Farfetch’d just barely made the cut, and Crocky or Cactus did not. So there you have it: Madame, a Pokemon which probably dated all the way to 1995 almost made its way into Gold and Silver, only to be replaced by Wooper at the last second. Poor Madame, and poor Farfetch’d: constantly being slighted, never truly being allowed to shine. (Madame art by @Raciebeep, as usual!) ID 320: SteelixAfter the complicated and convoluted Taaban and Madame, it’s nice to arrive at the simple, stable, rock-solid design of Steelix. Steelix is one of those Pokemon the team got right the first time. It had a great sprite in Spaceworld ’97, and a solid concept, and so it basically stayed identical the entire way through development. Still, there’s some interesting things that we can dig up about what it demonstrates about development around Era III. First of all, let’s be really clear: Steelix was 100% a Ken Sugimori design. The sprite has a lot of his signature shading and proportions, but even beyond that, Steelix matches the patterns we’ve seen with other Sugimori designs in the Korean Index. Most Sugimori designs stayed the same or were only minorly changed over the course of development: See Slowking, Crobat, or Girafarig for an example of this. Steelix, likewise, was unchanged from start to finish; it doesn’t even have any slight variations on its sprites in the Scratchpads, and its Silver sprite only has the tiniest variations to shading from SW99 to the final. Even the Crystal Sprite didn’t change at all. These characteristics, along with how generally well done the sprite is, convince me that no one but Sugimori could have made Steelix. Steelix fit an obvious niche, in retrospect. Since Generation II introduced the Dark and Steel types, one of the best ways to extend those innovations was to create an evolution of a Generation I Pokemon that had one of those new types. In the final Magnemite and Magneton are also Steel types, but that took awhile: Magneton and Magnemite were only changed to Steel type on July 30th, 1999, just a few weeks before Spaceworld ’99, while Steelix was a Steel-type from the beginning. And even when they became Steel-type, it just isn’t the same thing as having an entirely new evolution share that type. There were a lot of possible Pokemon without a third evolution which could have conceivably evolved into a Dark type: Fearow, Arbok, Golbat, Muk, maybe Persian all had the right flavor (and eventually the team decided that Umbreon would fill the role of a Gen I Dark-type relative). But there are far less opportunities in the original roster for Pokemon that could evolve into a Steel type. Conceivably Magneton could have evolved into a Steel type, before the team just decided to make the whole family Steel. Maybe Sandslash, Rhydon, or Cloyster could have evolved into a Steel type. But Onix, being a one-stage Rock Pokemon that already had a “strong defense” flavor made the most sense. Soon after creating Steelix, we’ll also see that the team decided that some of the bug Pokemon could also evolve into Steel Types, probably to buff their weak typing with something more defensive. Thus, Plux and Scizor were also created as ID 424 and ID 426 (which we’ll cover in the next article). Initially, Steelix was a simple upgrade to Onix: when Onix evolved, Steelix added twenty points to all of Onix’s stats and replaced its Rock type with its Steel type. All in all, that made it a pretty good evolution, but it was also fast: adding 20 to Onix’s speed gave Steelix a base speed of 90, on par with Pikachu, Golbat, Ponyta, and a lot of other speedy guys that Steelix has no business competing with. Eventually, the stats were revised to give Steelix a base speed of 30, which is actually 30 lower than Onix. That’s actually quite interesting to me: true, it makes sense that when Onix’s rock body is replaced with metal, it’d get weighed down even more. I like that flavor a lot. But that also means that there are niche situations where an Onix would be preferable to Steelix due to how fast it is. To my knowledge, no other evolution loses stats this drastically when it evolves, but I don’t know for sure; let me know in the comments. (Update: Many commenters have pointed out how common this is, especially since Generation VII! I guess I should've checked into this a little more!) I will say, I like this change to Steelix overall. It gives Steelix much more of its own identity. It’s now much more of a formidable beast than Onix is, but at the very least it’s a slow moving monster, weighed down by its own defensive armor. The main place most people remember Steelix from is that the fourth gym leader, Jasmine, who has an absolutely terrifying Steelix as her ace Pokemon. It’s one of the most powerful Pokemon the player’s faced up until that point in the game, and unless they have a strong Fire type Pokemon, it’s likely to be a roadblock for many players. That got me thinking: why does Jasmine have a new Generation II Pokemon like Steelix when so many of the other gym leaders only use old Kanto Pokemon? Faulkner, Bugsy, Morty, and Chuck all use only Generation I Pokemon, even though there were obvious new Pokemon for them to choose from. Faulker, the Flying Gym Leader, could have used a Hoothoot (it’s even on the route before you fight him), Bugsy, the Bug Gym Leader, a Scizor instead of his Scyther, Morty, who used Ghosts, could have used a Misdreavus, and the Fighting Gym Leader Chuck surely could have had a Hitmontop on his team. As for the other four, they still mostly only use Kanto Pokemon, with only one new Pokemon on each team: Steelix, Miltank, Piloswine, and Kingdra. This has always been a mystery: if the designers wanted to show off all the new designs they had created, why do the gym leaders so rarely show off these new Pokemon? It could be the case that the gym leaders’ lineups were developed earlier than the roster was completed and Game Freak only wanted to use Pokemon they knew weren’t liable to be cut. Miltank and Steelix, as Sugimori designs which more or less stayed identical throughout development, were probably always considered to be essential to the roster and thus safe to include in the Gym leaders’ teams. Kingdra and Piloswine, however, I have less of an explanation for. Kingdra was not a Sugimori design and it was redrawn at least once in development; in many ways, it strikes me as likely to make it to the final as Madame was. Piloswine, as well, doesn’t appear in the Pokedex until development restarted in 1999, but there are a lot of reasons to think that the team heavily invested in Piloswine being part of the final roster as soon as they developed it. It's inconclusive, but my guess is that all four of these Pokemon were ones the developers were pretty sure weren't going to get cut or drastically revised. This could be in part for balance reasons: if the Pokemon wasn’t completely finalized in terms of moves or stats, they probably didn’t want to accidentally make a Gym leader too challenging by changing how that Pokemon worked. Kingdra, as a stronger version of Seadra, might have been enough of a known quantity that the designers felt they could balance Clair even if Kingdra's stats changed a bit; maybe Piloswine was just too tempting to use for Pryce in a world without many good choices of Ice Pokemon. This theory also makes some sense of Chuck and Morty’s teams, for instance. We know that Hitmontop went through a lot of design changes and was completely revised only a few months before the game was complete, so they may not to have wanted to take a chance on putting Hitmontop in Chuck’s team in case he changed too drastically. Hitmontop’s also the only new Fighting type available, so if they didn’t want to use it, then Chuck would have to rely only on Kanto Pokemon. Likewise, as we’ve seen, Norowara and Misdreavus were in flux until the very last few months of development, so Morty probably couldn’t have used those on his team. But the theory fumbles with Bugsy, since Scizor is a Sugimori design, basically finalized even in Spaceworld ’97, and designed around the same time as Steelix. Maybe the team was just less sure that they’d need a Scyther evolution? Or maybe Bugsy didn’t use it for balance reasons, since his Scyther is already very difficult to beat in the final games. Likewise, I have no explanation for why Faulkner didn’t use a Hoothoot (or a Natu for that matter). The Elite Four and the Kanto gym leaders use a bit more of the new Pokemon, which makes sense, because they were designed and finalized much later and thus Game Freak would have had a better idea of which Pokemon were going to be on the final roster. On the other hand, the Rival uses a Sneasel even in the SW’99 build of the game, despite Sneasel being one of the most reworked Pokemon in the Pokedex. Maybe the decision to make Sneasel the Rival’s signature Pokemon was one of the reasons the development team was so persistent in revising Sneasel’s design? I really wish we had earlier versions of the Gym leader rosters from the time of SW’97; that would really help clear up some of what was going on here. Unfortunately, with what we have, all we can do is make conjectures like the above. I think the Gym leaders’ team selection has a lot to do with which Pokemon were the most finalized, but I can’t say that for sure. There are just too many outliers and strange situations. Anyway, this conversation has strayed away from Steelix. Steelix is cool. It’s a good design, and exactly the sort of exciting and attention grabbing design that a sequel should have. It was an inspired choice to give Jasmine a Steelix to really cement just how scary the new Steel type was, and I wish the designers had done this with more of the Gym Leaders. Given how rushed and manic the development of Gold and Silver were, I guess we’re lucky to have gotten Jasmine’s Steelix at all. ID 421 Blissey (Happiness) And we’re finally onto Blissey. Another evolution of an overlooked Generation I Pokemon, there honestly isn’t much to say about Blissey. Though it went through a few strange design concepts, in the end Blissey ended up just being a direct and obvious extension of Chansey’s aesthetic. Out of everything found in Period 3b, Blissey is probably the archetypical example of what this section of the Korean Index was about: extending the ideas of Generation I into new, cooler, evolutionary forms. In many ways, Blissey is in pretty much the same position as Madame. Both Chansey and Farfetch’d were rare and nearly unused in Generation I. At least Chansey appeared a little more frequently than Farfetch’d, showing up as a rare (and incredibly frustrating) encounter in the Safari Zone which will always run from you no matter what you do, and in the post-game Cerulean Cave. Still, to my knowledge no trainers in Generation I use Chansey, and given how hard it is to capture, Chansey was probably unused by the vast majority of players. Chansey’s a little bit easier to catch in Gold and Silver, though Chansey is limited to three routes in Kanto, making it, and Blissey, purely post-game Pokemon, another strange oddity in Gold and Silver. Still, Chansey was a one-stage Pokemon design basically begging to be expanded, and if Chansey was easier to catch in Gold and Silver, it was already going to be more prominent than in Red and Green. Why not give it an evolution that would service to reward players that went out of their way to raise Chansey? But what would an evolved Chansey even look like? The designers don’t seem to be sure. The SW’97 design is much stouter than Chansey, covered with extra feathery bits, and what looks like two eggs. It also has a heart on its head, which some people have commented looks a bit…rude. None of this strikes me as a good evolution for Chansey. It’s sprite gives the impression that it's shorter (or at least the same size) than Chansey, Blissey doesn’t look much like Chansey at all, the heart on top doesn’t have any analogue in Chansey’s design, and holding two eggs seems just about the most unimaginative way you could expand on Chansey. This strikes me as a real first draft, one the team thankfully moved away from. In June 1999, Blissey was redesigned again, this time to look a lot more like Chansey. Again, however, this redesign doesn’t work. First of all, it’s bright orange. I have no idea why the team decided that Blissey should be orange, and obviously they could easily change the palette, but I also think the color doesn’t match Chansey at all. The other problem is that this Blissey doesn’t look all that much like an evolution of Chansey. It basically looks the same, except it has an apron. That's rather fitting, as both Nurse Joy and the player's mother (Gen I and Gen 2) have one. It makes Chansey look more caring and motherly, but it's hardly visually striking.
That version of Blissey was then again reworked (probably by Nishida) through extensive edits to this sprite. The new designer seemed to like the color contrast this design got with the bright white apron, so they kept the idea of Blissey having a white bottom and a (thankfully) pink upper body. They also kept the curls of the June 1999 design, but redrew them a bit to look more natural on Blissey’s head, and gave it a cheerier expression (it’s name in Japanese, after all, is Happiness!). Most important, the new design philosophy seems to be “Chansey but make it look cooler.” So Nishida dialed Chansey’s girlieness up to eleven. Not only was Blissey cute and happy like Chansey, but now it has feathery frills and ruffles all over its body. It's honestly adorable: I know some people who think it’s too blunt a design after the more nuanced Chansey, but I personally think it fits well as exactly what I would expect from an evolved Chansey. Blissey’s stats are nothing special: it just gets an upgrade to the extremely high HP stat, and a bit to its Special Attack and Special Defense. But Blissey—and Chansey, since they share a moveset—do have a couple of interesting changes to their movesets from Generation I. In Spaceworld ’97, the designers gave Blissey and Chansey the move Pain Split, albeit learned at an extremely high level (73 for Chansey and 82 for Blissey!). In SW’97, the designers gave this move to three families: Norowara got it, because it was more or less the perfect move for a living voodoo doll that hurts itself to hurt its enemy, but the loveable and cute Jigglypuff and Chansey also got it, presumably because they had such a high HP they could use it effectively and not because it made any sense flavorwise. Thankfully, the team realized how out of character this move was, and by the final, it was only the signature move of Misdreavus. The other major change to their movelist was to make it so Chansey and Blissey learned Softboiled naturally, without a TM. Softboiled, if you don’t remember, was more or less a signature more of Chansey in Generation I, even if it didn’t it naturally: it either healed half of Chansey’s health in battle, or Chansey could siphon some of her life off and give it to another Pokemon outside of battle. Softboiled was one of the strangest moves in Generation I, because it was only learnable through TM 41, but the only Pokemon which was compatible with TM 41 (besides Mew, which could learn every TM) was Chansey. Why even bother? There’s probably some strange development reason why Softboiled was the way it was. It’s even odder because it’s one of only three non-HM moves that could be used outside of battle, the others being Dig and Teleport. Both of those moves did the same thing and brought you back to the last Pokecenter you visited; Softboiled had a unique effect which probably took a lot of programming despite being so obscure. Anyway, it’s another weird oddity from Generation I, and it makes far more sense to just let Chansey and Blissey learn the move. My guess, given how many times Blissey was redesigned and how it evolved from such an obscure Pokemon, is that Blissey was just as likely to be removed from the final roster as Madame, and only luck got us one and not the other. Blissey and Chansey only play a minor role in Gold and Silver, which makes Blissey feel yet again like another wasted design, sequestered to the post-game of Gold and Silver. At least the player would have been able to use Madame to defeat the Elite Four; Blissey is more or less an easter egg. I love Blissey and it's a Pokemon I've used a lot in later Generations, but personally, I think Wooper should've replaced this gal rather than Madame. And for all we know, it almost did. Imagine that in an alternate world, Chansey had to wait twenty years to get it’s evolution: the chivalrous Sir Bliss, wielding an egg shaped mace. |
AuthorMy name's Aaron George; I'm a Historian, and interested in Pokemon's development as a hobby. Contact me at @asmoranomardicodais Archives
July 2023
Categories |