The Poisoned Chocolates Case by Anthony Berkeley

poisoned_chocolatesSir Eustace Pennefather, a known womaniser, arrives at his club at 10:30 am to receive an complimentary box of chocolates through the post. Not being impressed by such a promotional stunt, he gives the chocolates to Graham Bendix, a man who, coincidentally, owes his wife some chocolates after losing a bet. As he shares them with his wife, he only eat a couple, as they have a distinctively unpleasant taste. She, however, eats a lot of them. He becomes sick. She becomes dead…

With the police reduced to looking for a random loony, Roger Sheringham convenes his Crime Circle – six individuals all who have an interest in crime. Each agree to investigate the crime for a week and give their solution to the rest of the Circle. Each of them has their own theory as to who the killer is. But which, if any, of them is correct?

One of the true classics of the Golden Age (so of course, this is the first time that I’ve read it), it is presented in a ground-breaking style. By concentrating almost entirely on the deductions of Sheringham and his colleagues – we get a little of Sheringham’s investigations as well – and the breaking down of each conclusion by the other members of the Circle, it’s almost as if we’re eavesdropping on a conversation between the writer of an unsatisfying mystery and a reader picking holes in it – I’m sure Rich from Past Offences could have such a conversation with Dame Agatha over The Murder Of Roger Ackroyd.

I do wonder if Berkeley was having a direct go at particular authors with each of the theories. It’s all written in an enjoyable style, so even if it is, it’s not a serious attack. I was caught out by the identity of the murderer – in fact I feel like I fell into the trap that Berkeley laid for me, given which of the sleuths plumped for the same option as me.

I’m not convinced, however, how much of a fair-play mystery it is. The solution makes sense but I’m not convinced if any reader would or could have worked it out. Not that this particularly bothered me as it was too much fun of a read.

Definitely a classic that needs to be read. Highly Recommended.


  1. I’m glad you enjoyed this one and managed to read it without being spoiled. It’s certainly a classic. I’m afraid I’m not a huge fan of Berkeley writing as Berkeley. I find them a bit dry, and not quite as funny as they need to be to sustain the gimmicks. Unlike our discussion of Christie’s gimmicks, I think Berkeley’s gimmicks regularly overshadow the books they’re in. If you read more of them I think you’ll find yourself solving them very early on, in an M Night Shyamalan sort of way. (You know there’s always going to be a major gimmick of some sort, so that drastically reduces the solution space. Of course here it’s much harder because even the false solutions are gimmicks of that sort!)

    It’s interesting that Sheringham is so flawed, and not in the standard “genius detective” way. He’s often very stupid and petty, and in books like Jumping Jenny he basically has no redeeming qualities at all. I admire Berkeley for taking such an unusual approach, but I’m not sure that’s actually the kind of detective I want to spend much time with. It didn’t help that the first one I read was The Layton Court Mystery, an atypical Berkeley that’s a direct parody of A A Milne’s absolutely dreadful (and definitely worth parodying) The Red House Mystery. It’s a clever enough mystery, but Sheringham spouts of the worst overt racism I’ve ever encountered in a Golden Age mystery, and I’m 90% sure we’re suppose to laugh along with it. This aspect of his character is dropped in later books, but it really soured me on him and Berkeley.


  2. The book definitely can’t be regarded as a fair-play mystery. I also found it dull.
    At first sight, it appears to be an expanded version of Berkeley’s short story The Avenging Chance. In fact, the first sentence of the short story reads as “When he was able to review it in perspective Roger Sheringham was inclined to think that the Poisoned Chocolate Case, as the papers called It, was perhaps the most perfectly planned murder he had ever encountered.”
    However, the correct solution of the novel is different from that of the short story. The solution of The Avenging Chance is one of the false solutions of the novel.


    • It’s light years off being the dullest Golden Age mystery I’ve read, but I do think the dialogue heavy nature of the gimmick takes its toll towards the end. Having one fewer detective might solve this, although it would unbalance things in different ways.

      I’m interested in this idea of it not being a fair play mystery. I don’t think I disagree, but you and the Doc have made two fairly definitive assessments about this that don’t quite jibe with my experience: “This book definitely can’t be regarded as a fair-play mystery” and “I’m not convinced if any reader would or could have worked it out.” I think those are two very different claims. Or rather, I think the book exists in the (sizeable) gap between those assertions, and that’s something I’m very interested in.

      I’d say I “worked it out” in the sense that, by halfway through, I was sure whodunnit. I use “sure” in the gambler’s sense here: I was sure enough that I’d have happily placed a large bet that my solution was correct. But that was based mostly on my experience with reading mysteries, my experience of Berkeley’s plotting style and extrapolating from what I’d seen of how the gimmick was playing out.

      Does the solution to a mystery HAVE to rely on in-universe clues? Does the reader have to play (or at least be able to play) the exact same game as the detective, using only the exact same information available to them? That seems like a reasonable definition of “fair play”, but it also seems an impossible challenge. After all, the reader exists outside the universe of the book, and solving mysteries is always going to require bringing in some kind of outside information. (If absolutely ALL of the information needed to solve a mystery was included in the novel itself, it would be a very convoluted logic problem, not a mystery.)

      This seems an especially thorny problem given the metafictional nature of The Poison Chocolates Case. It’s a mystery ABOUT mysteries, and even more artificial than most, so it seems odd to have to constrain our reasoning to the “real world” of the novel when that real world clearly doesn’t and can’t exist. But given that all readers have different experience, how can a mystery be fair if the information needed to solve it comes from outside the book? I don’t know how to resolve the contradictions that arise here, but they interest me far more than any stipulations about secret passages and twins.


      • Certainly my intention was to imply that while the information might well be there for the reader to spot, it’s so well hidden that the likelihood of anyone spotting it is so low to make it effectively solveable. Maybe we should develop the phrase “unfairplay mystery”. I recall Squire Throwleigh’s Heir by Michael Jecks – there’s a tiny something about the crime that even the sleuth doesn’t spot until after the killer is caught that would have given them away. Similarly in Shatter The Bones by Stuart MacBride (although that one is in plain sight throughout the book). In both cases, you can make a good case that these are the only real clues in some longish books. On the other hand, there was much more of an “oh…” moment when they are revealed with those than in this one. Does that make them fair play? In my opinion, trying to set a precise definition as “fair-play” is almost as difficult (and unnecessary) as trying to define the “Golden Age Style”. Pity the fool who might try and do that 🙂
        Back to this one though, yes, it’s far from being dull, although yes, maybe one less sleuth might have worked.


      • Sure. Precise definitions aren’t possible or useful for such a nebulous concept. But I do think there’s profit in investigating the phenomenon where it can be clear whodunnit even though there aren’t many (or even any) in-world clues. I’m perfectly happy to believe there isn’t a single piece of evidence in the traditional sense that shows whodunnit in this book (it’s a long time since I read it, and certainly nothing springs to mind). But at the same time, I think the identity of the murderer is both inevitable and fairly obvious. The tension between these two assertions seems worth exploring, both in a philosophical sense and as an inroad into more practical stuff to help writing technique (such as pacing and audience manipulation).

        Perhaps a useful distinction would be between something like “solvable” and “gettable”, where a mystery is solvable if you can reach the solution using clues and evidence from within the story, and gettable if you can reach it using other means. To take an extreme example, I think “What Happened At Hazelwood” by Michael Innes is both technically fair and also completely unsolvable – there are a lot of clues, but you need to have such a strange imagination to work out the solution that really the only person who could be expected to solve it on those terms is Innes himself! But it’s also one of the most “gettable” mysteries I’ve read, because I was pretty sure who the murderer was simply from reading the contents page (it’s more complicated and interesting than a giveaway chapter title, but it’s still a bit of a shame, given how much effort went into the construction of it).

        The distinction is important, because if a mystery is too “gettable” then it undermines efforts to make it “solvable”, and solving a mystery is by far the more satisfying experience. “An Instance of the Fingerpost” is a good example. It’s an incredibly clever historical mystery with a lot of well hidden clues and narrative tricks that it’s perfectly possible to pick your way through with a bit of thought. But that was completely undermined because I “got” who the murderer was going to be as soon as I knew what the structure of the book was (using almost exactly the same reasoning as for What Happened at Hazelwood). Authors may want to be aware of these shortcuts so they can try and conceal them better. Conversely, now that fair play puzzle mysteries are less fashionable, authors may want to study how to make a mystery “gettable” without being “solvable”, because that can make for a satisfying conclusion while avoiding reliance on standard types of clues.


      • I’ve wondered about this sort of thing, too. There is often at least one really obvious suspect in a mystery, but I end up thinking to myself: “It can’t be him because it would be too obvious.” But in real-life cases the obvious suspect is usually the guilty party, whereas in mysteries the obvious suspect is usually a red herring. This sort of meta-level reasoning feels like cheating to me, except in the rare case of a story that admits it’s fiction within the story itself–such as John Dickson Carr’s “The Three Coffins (The Hollow Man)” or some of Edmund Crispin’s books. Nonetheless, I’ve read too many mysteries not to think this way when I read them.


      • The thing that I most liked about this one was that I immediately leapt to the wrong suspect due to it having a structure that Dame Agatha overused, and so I missed the real killer by a mile. I’m sure that was done deliberately by the author.


      • However, in the short story Avenging Chance you would have leapt to the correct suspect.


  3. definitely an expansion of AVENGING CHANCE and clearly a bit of a spoof, but a hugely entertaining one – a firm favourite, always been part of my top 100 and I can;t se eit ever slipping out.


  4. Incidentally, the short story The Avenging Chance is included in the book Capital Crimes: London Mysteries edited by Martin Edwards.


  5. The latest edition of the book published by the British Library has 2 extra solutions by Christianna Brand and Martin Edwards.
    That is my major complaint against the book. So many correct solutions are possible !


  6. Just bought the British Library edition over my Kindle, despite already owning a hard copy – for the sake of the two additional resolutions. 🙂 Currently re-reading it.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.