I just now finished today’s Listener crossword, “Merchandise” by Adam. After filling in the grid correctly, you still have to find a quotation that is “embedded in a regular way” in the grid. I spent a while staring at the completed grid and trying out different ideas — looking for words forward, backward, and diagonally in word-search style, trying every second or third or fourth letter, trying out zigzag paths and knight’s tours and so on — before I finally found the quotation, but I have it now. It’s a very familiar quotation to me — the first words in it I spotted were the third through fifth, and as soon as I found half of the sixth word, the whole line sprang into my head.

1 Down and 16 Across are two more examples of a kind of clue I dislike, where the wordplay makes use of an obvious form of the very same word, like (this is a made-up example) clueing BELOVED as “Cherished one showing passion in bed” (that is, BELOVED composed of LOVE inside BED). 22 Across is of this sort, too, though not quite as blatantly.

You see this kind of clue a lot in British cryptics, but rarely in American ones. Is it a cultural thing? Does this sort of thing just not register to the British puzzle-solving sensibility as being lame?

Still, the puzzle has a number of nice clues, and it was a pleasant challenge. And I’m too busy with working on the play this weekend to be knocking my head against the Listener puzzle all day today, so I’m just as happy to have a moderately easy one this week.

Later: This posting has become the most popular one on this blog in a long time! It seems that a lot of solvers are not having any success in finding the quotation and are searching the Internet for help. I don’t like to give things away, so I’ll just say that the phrase “in a regular way” in the introduction to the puzzle is straightforward and accurate.

I’ve read some snarking online to the effect that the manner in which the quotation is hidden should have been indicated by the theme itself rather than by a phrase in the introduction. I’m not sure I see why; it’s certainly fun when it works out that way, but it doesn’t seem to me to be an unwelcome change of pace to have a puzzle that ends with something of a treasure hunt like this one does.

This week’s Listener puzzle, “Annual Turnover”, was a little easier than average in difficulty, and a good thing for me, too, as I had a lot of rewriting of the play to do this weekend. The 13-letter theme word at 28 Across is a familiar one to me, as I’ve seen it used in word puzzles a number of times, and I got it from only three or four crossing letters; that helped a lot in finishing the puzzle. I also guessed at what the pattern of clashing letters was after I had only three or four of them, and I feel fairly sure I’ve seen in an American crossword somewhere or other the same device of replacing the letters in certain squares with the same symbol used here, resulting (if my memory is right) in the very same pattern. Maybe in a crossword in Games magazine?

As I copied the answers onto a fresh printout of the grid, I noticed that the setter could have removed several of the bars to make some of the words longer and give more crossings. Now, Ximenes developed the practice of giving every word in a bar-style cryptic crossword at least one unchecked letter (that is, a letter used only in one direction and not crossed by a word in the other direction), so that none of the answers would completely fall into place without the solver having to solve the clue. But he intended that for plain bar-style cryptics, and I don’t think it applies so well to a novelty cryptic like this one, especially where the puzzle’s gimmick interferes quite badly with the usual help you expect to get from crossing letters. Giving a few additional crossings where you can seems just sporting to me, and anyway the longer words are usually more interesting for the solver to find. Besides, there are already entries in the grid here with all their letters checked — 3 Down, 20 Across, and so on. So, unless I’m missing something, I think the setter would have done better to give the extra help and give more crossings where he or she could.

To wit: If the three-letter entries at 15A and 47A were extended to four letters, symmetry would be preserved and both would still be valid entries (that is, either words or phrases found in Chambers). Similarly, if you erased a couple of bars to extend the symmetrical pair of five-letter entries at 1D and 42D to six letters each, one of these would still be a valid entry, and the other could easily be changed to any of several possible entries because of the multiple possibilities for its second, third, fourth, and fifth letters. (Another plus: Several of the possible entries would have been more interesting than the one actually used.) Again, if 14D and 34D were extended by one square (toward the edges of the grid), one of the longer entries that resulted would already be valid, and the other could be turned into a valid entry just by changing one of its unchecked letters. Finally, running down the middle column, above and below 19D, are two perfectly good potential three-letter entries that are instead divided by bars into individual unchecked letters.

Still, all in all a pleasant puzzle, not terribly remarkable or surprising or difficult, but fun to solve.

And for a numerical puzzle, too. I solved this even more quickly than the puzzle of two weeks ago: during my lunch break, in well under half an hour!

Unless I’m really missing something, there isn’t anything hard about the puzzle at all. The entries in the grid are all two- or three-digit primes whose digits are all odd. It doesn’t take all that long to list them all (there are 54), make sublists of the ones that match the given categories (two-digit twin primes: 11 & 13, 17 & 19, and 71 & 73; three-digit palindromes: 131, 151, 191, 313, 353, 757, 797, 919; and so on), and then try out the not-very-many possible combinations until you find the ones that fit together in the right places in the grid. I was never stuck for what to do next at any point; I was limited more by how fast I could write than how fast I could make deductions.

There is one neat touch: The grid, a 9 × 9 square, turns out to have exactly 54 entries, which means it uses each number exactly once. It’s a nice piece of construction to pack exactly that set of numbers into a perfect square, and have the resulting grid be symmetrical at that. But the clues give way more information than is needed to figure out how to fill the grid.