Listen With Others

I like grids which portray symmetrical themes and twelve is a nice number. It is pure chance that this puzzle follows CHAT. This is because CHAT was held back several years by the editors since a puzzle on the same theme had been published in 2012.

The number twelve immediately suggested the signs of the Zodiac in a circle placed around the Earth. The puzzle was constructed in June 2016 a few months after Mynot’s “Stomach” (4380) which had the sun and the planets on a diagonal with their names replaced by symbols, and I could implement the signs of the Zodiac in a similar manner in an unnumbered grid so that the locations of the symbols would not be revealed. Because the names were more complex, they could not be split into two parts and would need to be anagrammed. My submitted preamble was much longer than the published version and contained an example of two crossing answers producing the word ASTERISK; in particular it said that there was at most one thematic cell in each entry.

Normally in an unnumbered puzzle the solver will try and fit the longest words in the grid in order to make a start, but this would not work in this case. The grid with 180⁰ symmetry was constructed so that only just over a third of the cells were both checked and had 90⁰ symmetry and could therefore contain a thematic entry (see shaded cells in the diagram). The grid contained four 3-long entries, but only two 3-long answers were clued, so at least two of these entries must contain a thematic cell. Checking possible 8-long answers intersecting with 3-long answers would show that none of the thematic cells were in the perimeter and lead to the discovery of PISCES and VIRGO, after which completion of the grid is straightforward.

The original grid had a vertical bar-line to the right of LIBRA [at 6 o’clock] which forced the down entry to be ALBIAN, a word which is not in Chambers. When I reviewed the puzzle before submission, I realised that this bar-line could be moved to the left of LIBRA and the down entry changed to ARABIA leading to small changes in the SE corner. The opposite word SERIES would still work with a few changes in the NE corner. I have often found that it is worthwhile to tweak a grid by moving a few bar-lines, or, even omitting some, to produce a grid with “nicer” words.
 

Postscript

I have now received a large number of comments from solvers, which were, in almost all cases, very positive. Thank you everybody for sending them. Two things stood out – most people solved almost all the clues before making any entries and people were impressed by the construction of the grid.

Only one solver counted the length of the answers before solving any clues and found that there were 53 letters absorbed by the symbols in addition to the letters which would normally have been in the cells. The number of symbols was either 4n or 4n+1, and a guess of twelve symbols would produce 77 letters for 12 words with a very plausible average length of 6.4 letters. (I would normally tabulate the lengths in puzzles where the answer lengths differ from the entry lengths and I would have expected other solvers to do this sort of analysis.) This solver also noticed that the total number of letters in the twelve signs of the Zodiac was 77 letters. The other thing that I expected an alert solver to do was to shade cells which show 90⁰ symmetry and were checked – this gives a clear suggestion of a circular shape.

 The other property that was designed into the grid was four 3-long entries, but only two 3-long answers; hence at least two (and in fact all four) 3-long entries had to contain thematic cells. The clues for the 3-long answers were easy and three of the four longest answers were anagrams, so that most solvers should have recovered these answers fairly quickly. After solving half the clues, an alert solver should have been able to place LAP with INACCESSIBLE to produce PISCES and VAN with RADIOGRAPHY to produce VIRGO, which would have revealed the theme. Most solvers solved virtually all the clues before attempting to start to fill the grid.

The construction of the grid took less than two hours. An initial grid was constructed so that only a third of the cells could be thematic; in particular there were only four possible positions in the perimeter. Seven of the Zodiac signs could be split into one letter and the remaining letters; the remaining signs were split into two roughly equal halves and TEA was used to expand them into possible words. Seventeen “words” were fitted into the correct places in the grid with the seven single letters replacing the clashes. I then used Qxw to obtain an initial grid which was then improved by making some small changes to get rid of some awkward words.