The Tokyo Zodiac Murders (1981) by Soji Shimada

On January 31th, 2019, Pushkin Vertigo is finally going to release the long anticipated translation of Soji Shimada's Naname yashiki no hanzai (Murder in the Crooked House, 1981), his second novel to appear in English, which Ho-Ling Wong characterized as a detective story in the vein of "the classics of the good old age" – complete with an entirely original locked room-trick. So I decided to commemorate this upcoming release by rereading his bloody tour-de-force.

Soji Shimada is considered to be "the doyen of the Japanese form of Golden Age detective fiction," known in Japan as shin honkaku, which can be traced back to the publication of his debut novel, Senseijutsu satsujinjiken (The Tokyo Zodiac Murders, 1981). A horrific, labyrinthine jigsaw-puzzle involving severed body parts and a seemingly impossible murder.

The Tokyo Zodiac Murders is the case of "the Umezawa serial murders" which took place in Japan in 1936. One of the most elusive mysteries in the annals of crime and the story opens more than forty years later, but is prefaced with the last will and testament of Heikichi Umezawa – a mentally unbalanced artist with a deranged plan. Umezawa is obsessed with creating "the perfect woman of supreme beauty," Azoth, who wants to bring into existence. A dark, demented fantasy requiring the body parts of "six virgins of different zodiacal signs" and fate has handed him the women he needs on a silver platter. Namely his daughters and nieces!

Umezawa begins to plot the genocide of his own relatives and strictly follows the rules of alchemy, in correspondence with the astrological signs of the victims, but he's inexplicably murdered before he could carry out his plan. Umezawa had installed iron bars over the windows and skylights. The door of the studio was "a Western-style, single-panel door" that opened outwards with a bar to secure it from the inside, which was in place when Umezawa was murdered. So how did his murderer enter or leave the studio?

However, this is only the beginning of what would become the "genocide of the Umezawa family" and "consist of three separate cases."

The second case is the murder of Umezawa's stepdaughter, Kazue Kanemoto, who was raped, beaten to death and her rooms were ransacked, which made it look like "a run-of-the-mill murder" to the police – probably by a burglar. You would think rape was the one crime you can't possibly use in an traditionally-structured, plot-driven detective novel, but Shimada actually shaped a disgusting rape-murder into an important piece of the puzzle. Finally, there are "the Azoth multiple murders." Someone had carried out the dead artist's instructions to the letter.

Over a one-year period, the six mutilated, often badly decomposed bodies of Umezawa's daughters and nieces are found buried all over Japan. The bodies were found in places corresponding with the metallic elements specified in Umezawa's notes. All of the victim's were missing various body parts. This added one last, tantalizing question to the case: was the murderer successful in "creating the monster," Azoth, and where's this patchwork body buried?

The Tokyo Zodiac Murders never stopped to capture the imagination of the public and it periodically became "a fad to try to solve the mystery," but the case remained unsolved for more than forty years. Until one day, Kiyoshi Mitarai, an astrologer, fortune-teller and self-styled detective, received a client with a letter from her dead father. The letter throws a new light on the baffling aspect of the disposal of the bodied, but the woman wants Mitarai to unearth the whole truth. And clear her father's reputation.

What follows is arguably one of the most composed, cerebral detective stories ever written. The Tokyo Zodiac Murders has a plot involving dismembered bodies reminiscent of Resurrection (1999) and can go toe-to-toe with Michael Slade's Ripper (1994) when it comes to gory murders, but the story takes place forty years after the killings. So the investigation is purely focused on solving the puzzles.

Mitarai spends most of the first half listening to the narrator, Kazumi Ishioka, who's "a huge fan of mysteries" and gives him all the details of the case. During these parts, Mitarai comes up with good, but wrong, explanation for the locked studio that the police took a month to work out – helped by letter-writing armchair detectives of the public. A false solution that was obviously modeled on a very well-known short story by an English mystery writer. Once he has been filled on all the details, Mitarai takes Ishioka to speak with as many people as possible who were linked to the murders. And are still alive. The only real hitch they have in their investigation is that circumstances imposes a deadline on them, but the focal point remains piecing together all of the pieces.

My copy of The Tokyo Zodiac Murders is the original 2005 hardcover Stone Bridge edition and is littered with crime-scene diagrams, maps, charts and illustrations, which helped selling the historical aspect of the 1936 murders. There are enough maps, charts and illustrations to give you the idea of a murder-mystery presented as a case-file story. I appreciated the story had not one, but two, challenges to the readers that fitted the pure puzzle aspect of the plot.

There are, however, one or two blemishes. Firstly, the solution to the problem of the locked studio is fairly routine and surprisingly uninspired. I remember being slightly more impressed with it the first time around, but then again, there are nearly a thousand locked room stories between my first and second read. So I probably have become a bit harder to impress when it comes to the impossible crime genre. Secondly, the motive is weakly handled and tacked on at the end as an afterthought, which would explain the epilogue because it tried really hard to give the murderer the motivation need to carry out such a risky and insane plan – only it felt like it came way too late in the game. It actually came when the game was already over. The reader should have given hints to the motive a lot earlier in the story.

Nonetheless, The Tokyo Zodiac Murders is an impressive debut and the central puzzle of the plot, tying all three cases together, is nothing short of ingenious. Stuff of classics! On top of that, the cleverly hidden murderer is as skillfully handled as John Dickson Carr's tight-rope act in The Plague Court Murders (1934; published as by "Carter Dickson"). Personally, I can't think of a bigger compliment to give to a writer of traditionally-styled detective novels than that.

I'm looking forward to the release of The Murder in the Crooked House and I'll probably be reading, or rereading, another Japanese locked room mystery for my next post. I just have to decide which one.


  1. Though no one ever mentions it this mystery novel is inspired by a logic puzzle I remember from my elementary school days. It involves making several geometric shapes using pieces. But it’s not a Tangram puzzle because in this one you start out with a certain number of shapes and the puzzle is to create the same number plus one more. I wish I could remember the book in which it appears.

    1. Well, it kinda gives away the whole crux of the central puzzle of the plot. So perhaps that's why nobody ever mentions it. ;)

  2. I was very lucky to have read this novel before reading "the Kindaichi version". I remember I was so impressed by this novel, my first thought was to bring it to my brother who's also like to read detective stories (in fact, i was introduced to Conan and Kindaichi by him).

    His reaction when I urged him to read this novel was, unfortunately, not to my expectation. It was underwhelming.

    Looking back at that moment, probably the cover of the Indonesian version gave the solution away to him (i'm not kidding, if Mitarai saw the cover, he may reached the solution there and then). And knowing that my brother read Kindaichi, maybe that reminded him of "the Kindaichi version" with the same solution as the cover suggests.

    Although, I was not impressed by the novel's locked room problem, this is clearly one of my favorite detective novel.

    1. "I was very lucky to have read this novel before reading "the Kindaichi version". "

      Same here! I would have been so pissed had I read The Mummy's Curse before The Tokyo Zodiac Murders. Kanari always cribbed ideas left and right for his plots, but The Mummy's Curse was borderline plagiarism. Basically an abridged version of Shimada's plot.

      This is why I'm afraid that, one day, an incredible Japanese mystery will get translated only to discover Kanari had "borrowed" from it for House of Wax.

      "...in fact, i was introduced to Conan and Kindaichi by him"

      Now that's what I call good guidance from an older brother!

  3. I thought the date was moved from January to July for the secondo one.

    Thanks for the notification! I purchased it yesterday (Along with Higashino's latest) and voilla it's already on the way!

    1. This review was written last year and the scheduled release date then still was January, 31th.

      Well, I just checked and there appears to be two different release dates: UK (January) and US (June). So we'll have to wait and see.

  4. I've pre-ordered The Murder in the Crooked House - it comes out June 25th in the US. This one sounds fascinating. It doesn't sound like you were put off by the gore.

    1. Gore doesn't bother me at all and a dismembered body usually serves a purpose in Japanese detective story (i.e. corpse-puzzles). They came up with some completely new and original locked room-tricks by playing around with severed body parts.


    Is is blasphemy if I admit I didn't care much for this when I read it? The translation came off as stilted and the pacing in the second half is poor. Once we get the time limit set up Mitarai spends the rest of the book bumbling around before the reveal. That's an issue with these stories that look at older mysteries, I feel, especially here where all of the investigation doesn't really accomplish anything beyond giving out red herrings.

    I think my issue was that I thought that the locked room would play more of a role in the plot and have a more interesting solution. The false one had more creativity to it, I felt. I think that the central trick in the book is good, but sadly I spoiled it by accident so it lacked the needed oomph. And the killer felt...thrown in, I guess. They needed to make more of an appearance earlier, I feel.

    That being said I'm still looking forward to Murder in the Slanted Mansion. Shimada looks like he has imagination and ingenuity in spades from what I've read about him on Ho-Ling's blog, and I'm happy to give him another chance (also "The Executive Who Lost His Mind" was ridiculous in all the best ways, so that helps.)

    1. You're being a little blasphemous, but you make a valid point about the locked room and false solution, because the book has always been billed as a classic impossible crime. Technically, it would have been better if Shimada had found a way to swap around the false and true explanation to the locked room murder, but that would have been a tricky thing to do with the overall solution in mind.

      Murder in the Crooked House is a much more conventional detective story, especially when compared to The Tokyo Zodiac Murders, with a genuinely original locked room-trick. You'll probably like it a lot more.

      However, you can get a newfound appreciation for The Tokyo Zodiac Murders by the simply reading The Mummy's Curse from the Kindaichi series.