[Avian is a collaboration between Brock (Andrew Varney) and Hedge-sparrow (Rob Pinnock)]

AV: When vetting my debut Listener puzzle, Not a Black and White Decision (NABAWD), Derek Arthur commented that it looked like the first of a series. This puzzle was originally going to be NABAWD2. My notes from November 2018 suggest a similar arrangement, with clashes in a R/L format leading to WATERSHIP (Down) and some of the characters from the book appearing, along with SANDLEFORD and/or WARREN being destroyed. A second note refines this to show HOUSING CRISIS as a central band. At the very end of 2019, I approached Hedge-sparrow, a setter whose puzzles on not dissimilar themes I much admired.

RP: Andrew’s approach to me was the first time I had been offered the opportunity of participating in creating a collaborative puzzle. I wasn’t quite sure what to expect and, although Andrew explained what he had in mind and pointed me towards his original NABAWD puzzle, it took a while for me to begin to get the hang of collaborative puzzle creation. I remember I quickly created a “draft” puzzle which was not quite in line with Andrew’s concept, but we soon began to work very constructively (even though Andrew had to put up with my “pencil and paper” approach to puzzle construction and deal with lots of photographs of trial grids I sent him!) One of two relevant dates we had in mind for publication of the puzzle was the 50^th anniversary of the original publication of Watership Down (November 1972): we hoped that we would be successful in creating a suitable puzzle in time to “reserve” that.

AV: I was impressed that Rob had come up with a grid in about a week after almost 18 months of me sitting on it! His greater experience in setting showed through and he introduced some ideas that I’d not considered, which ultimately shaped the direction the puzzle took. That included the primary concept of “characters” being displaced from the left-hand side of the grid to the right-hand side, thereby overcoming the “housing crisis” rather than necessarily focussing on clashes.

RP: Early trials focused on trying different arrangements and positions for the phrases HOUSING CRISIS, SANDLEFORD WARREN, and WATERSHIP DOWN, as well as considering how to include the names of at least some of the rabbits. In kicking around various anagrams of SANDLEFORD WARREN between us, I discovered LAND FOR WANDERERS, which suggested the idea that the “characters” of SANDLEFORD WARREN could be expunged from the LHS (representing the destruction of Sandleford Warren) and relocated on the RHS (representing the creation of the new warren on Watership Down): in this concept, the term “characters” could then suggest either the actual letters of the phrases, or the rabbits in the story. After further discussion we concluded that, to make the puzzle solvable, the letters LAND FOR WANDERERS appearing in entries on the RHS should: (i) be clearly indicated by the wordplay of clues for those entries; (ii) if possible, appear in clue order. With 16 letters in the phrase, point (ii) seemed to be very difficult to achieve in half a grid: however, Andrew had the brainwave of splitting the phrase into LAND FOR appearing in across entries, and WANDERERS appearing in down entries, and “doubling up” the use of the letters AND appearing in both parts of the phrase. This trick enabled him to come up with a viable RHS with the letters of the phrase appearing in the right order. I had the rather easier task of creating a LHS with the letters of the phrase randomly removed from answers before entry: however, in creating both these grid-halves, we also had to consider the names of the rabbits – which to include, and how.

AV: As Rob mentions, I noticed that AND was repeated in LAND and WANDERERS, which finally enabled us to overcome the “gridlock” by allowing them to do double-duty. Even so, it was still a significant challenge to keep the grid symmetrical apart from the central band. At this point we had the four names HAZEL, ACORN, SILVER and FIVER in the bag, but having re-acquainted ourselves thoroughly with the book in preparation, it felt inadequate not to include BIGWIG and PIPKIN. While I use compiling software as a tool, it was primarily a guide in seeing what was possible. Rob’s coup was working in THERB[L]IG and TH[E] PIP with his pencil and paper to be able to include PIPKIN and BIGWIG in the final grid. Which brings us onto the clues.

AV: We allocated the clues alternately to one or the other of us, then discussing further refinements via SMS, email and a couple of Zoom calls. Writing good right/left clues is not easy, one secret (as per Don Manley) being to disguise the join between the two halves while maintaining a viable surface reading. Much time was spent on this and almost as much on trimming them. We competed to see who could achieve the shortest average clue length, neither of us being renowned for our concision. We had given ourselves an additional constraint: in Rob’s words “we did not want to leave any of the characters behind”, yet of the original eleven who left Sandleford to arrive at Watership, only six were in the grid.

RP: The five escapees not included were HAWKBIT, BLACKBERRY, DANDELION, BUCKTHORN and SPEEDWELL. Since these names could not realistically be incorporated in the grid, we considered ways to include them as part of the clues. With the exception of DANDELION, all the names split into two real words, which suggested that we could try to fit half of a name in each half of a double clue. Using “extra” words to be removed before solving would be too obvious, losing the PDM of finding the names in the grid, so we decided to incorporate them as essential parts of the wordplay. The most problematic name was DANDELION, but Andrew noticed the possibility of splitting it as “D AND E” and “LION”, and this trick enabled us to create a (somewhat convoluted) clue for the entry STEAK DIANE. We gave no hint that these extra names were “hidden” in the clues: they were there as an “Easter egg” for solvers to spot, though we’re not sure if anyone did.

AV: I mooted asking solvers to highlight all eleven characters, but Rob advised against it, bearing in mind some prior controversy related to leporine extra-grid highlighting. The editors kindly allowed us to keep the feature in the puzzle, despite it not being needed for the solution. The final editing stage seemed more involved than usual as we worked with the editors to balance surface reading, accuracy and concision in many of the clues. The title also changed at this point, since the puzzle was no longer obviously a NABAWD sequel. I suggested NEW ARRANGEMENTS, i.e. ‘NEW ARRS.’ (see Grove), reflecting puzzle mechanisms, story plot and leading to WARRENS with N, E removed from the left half and replaced on the right in a different order. To align better with Chambers, Rob recommended the improvement NEW ARRIVALS. As in the Watership Down story itself, the puzzle construction was quite an adventure, involving close collaboration between us and with gratefully-received help from others on the journey. Thank you particularly to the editors for this assistance and solvers for kind words of welcome following the puzzle’s publication.

Since reading Behind The Scenes at the Museum, her Whitbread-prize winning debut, I have been a fan of Kate Atkinson. Indeed, she is probably my favourite living author.

When musing over her output, it struck me how many titles were rendered cryptically – almost as if Kate Atkinson might be another crossword fan. And Life After Life, her first Costa Book Awards winner, lent itself to a title. Hence Listener 4735, which, delightfully for me, was published whilst I was reading her latest book, Shrines of Gaiety – and relishing it.

Huge thanks to Hedge-sparrow, Vismut, Dysart and the Listener editors, who all provided improvements to my clues.

Apologies for the double definition clue for ‘mossie’ – it should have read South African birds or sparrows, not swallows.

I submitted Spirit Time with a degree of doubt, since the Listener Guidance for Setters states that: “Themes that revolve round a living person are also discouraged because of the risk of death or dishonour at the time of publication.”

Upon acceptance, I assumed that they, like I, thought dishonour low risk in the case of Ms Atkinson. Like many of her characters, she may yet surprise us!

The inspirations for the puzzle were the two definitions of space (the other being “intervening distance”) and the definition of DEFINITION as “sharpness of outline, visual clarity”. I really liked the “material bodies” definition, found by chance. I understand that its vocabulary owes something to the philosophical literature, but the lexicographer should take credit, all the more so because you’re little the wiser after reading it.

There was a second sequence, the submitted means of confirming SPACE, but it depended on a “close but no cigar” pun I lost faith in before submission, and which the editors declined, subsituting the initial letters scheme. It’s there, although not organised neatly, and using the same values of abcde.

An 11×13 grid would have been ideal with the words in the external rows breaking THATINWHICH/HAVE EXTENSION, but the remainder of the definition would not have extended into them as they did; the fourteen letters of MATERIAL BODIES floating about in the grid would have been a very different proposition.

The overlapping of the messages was pure serendipity, which I gratefully hung on to. Naturally, I looked into whether applying the sequence in reverse column order, reverse row order, starting from the bottom right hand corner, etc. produced any further possibilities. I recovered most of my senses eventually.

My main regret was the possibility of a correct solution for some by simply guessing at it, having identified the definition, and being unwilling unravel the rest. Feedback suggests that this was not the case, however, with at least one doughty solver actively preferring to see the whole thing through.

The puzzle took a considerable amount of time to put together. Thanks to the Listener editors, the solver who tested it and gave pertinent feedback, QXW, and all solvers providing their feedback too — very welcome, and all points have been carefully noted.

I have learnt a lot in setting my first crossword; in particular, the feedback from the vetters Shane and Roger has been invaluable, but — as I’m sure many first-time setters find — the very first lesson is not to overstretch oneself! I was originally going to have the three Forgotten Things as anagrams to be unscrambled and Phlebas just to appear in the grid, but my insistence on the results being real words meant that finding so very many words which would work with both letters proved impossible — even Phlebas himself was only jumbled a little bit in the end!

One thing I was triply careful of was to make sure there was an entry point for a jigsaw-type puzzle with no clue numbers, so that some actual crossing words could be used to solve the crossword! “ABASHMENT” was my main solution to that — the only 9-letter answer, knowing from the clue order it must start somewhere near the beginning of the alphabet, and a relatively easy clue (and, at the risk of blowing my own trumpet, my favourite clue of the whole puzzle in terms of elegance!), but the four 7-letters in combination with the two 10s and the one actually clued 13 were also possible entry points. A friend of mine reported that he got in from the bottom left, whilst my father (who introduced me to the Listener many years ago, and with whom I was exceptionally smug when telling him why I couldn’t help with this one!) started with the 9 and with “CHEATING ON” (another clue whose surface reading I liked).

A lesson I have taken particularly to heart (the reason for most of Shane’s changes to my clues!) is never to imply direct equivalence if the wordplay doesn’t result in the light!  For example, my clue “Misguidedly write clue, revealing handbag” implied that the wordplayed answer with an extra letter e.g. “RETICULE + [W]” and the light “RETICULE” were identical. Shane’s correction — “Write clue about handbag” — doesn’t make any such claim. That was only one of many! Shane was kind enough to preserve something close to the surface reading wherever possible.

There were some much more significant changes, of course, which may interest prospective setters — my original clue for “UNTOLD + [O]” was “Indescribable treasure held back by United Nations diplomat at first”, but Shane thought that “treasure” and “loot” were not sufficiently synonymous; my original clue for “PLAN + [A]” was “Article after friend’s scheme”, which I liked for its deliberate ambiguity between the possessive and the abbreviated “is”, but that “is” was its downfall, implying — as I mentioned earlier — direct equivalence! (As an aside, I’d have been perfectly happy as a solver to work with that, which is why it came as quite a surprise to me that it didn’t pass muster! I might even suggest that the vetters add that to their “Clueing errors” page!)

People who solve my debut puzzle and find it to be a Waste Land reference which avoids either the first or last lines of the work may suspect that I am a particular student of Eliot… but in fact there’s a rather roundabout route to this theme. My first encounter with Phlebas the Phoenician was in the title of Iain M Banks’ first science-fiction novel, “Consider Phlebas”, without which I might not have read any of the Waste Land at all. The reason that Death by Water and its unfortunate, for want of a better word, protagonist in particular is memorable to me is a consequence of those arcane workings peculiar to groups of friends by which particular phrases or references become overused in-jokes! That the centenary of the Waste Land’s publication was coming up in the near-ish future (I compiled this around Christmas 2020) was a lucky coincidence, albeit one I was perfectly happy to make use of in asking for a particular date for my crossword. “Remembrance” did seem the perfect title, though — the surface theme of forgetfulness and the stanza’s theme of mortality combined in one word.

There were other possibilities for my first submission, of course, over which T S Eliot won out. I’m actually a mathematician, and the only reason I didn’t start with a maths puzzle — always my favourites! — is that, at a rate of four per year, I would be far less likely to get published than the average of forty-eight word-based crosswords per year. There were also Latin and Russian sources I considered which I was forced to discard because I didn’t want to be reliant on a particular translation.

Time, I think, to start considering my next submission — I’ve got an idea for a clue/entry method I haven’t seen yet, but I have no idea whether it will survive contact with reality…

I first came up with this idea in the year 2000, when my father (Aedites) first started creating Listener crosswords. At that time, I created a list of possible 3-letter homophones for letters of the alphabet, and a small 4×4 mini-crossword with 4 symmetrical barlines. The lack of A, E, and various consonants made the 4 barlines necessary. Back then, my father then tried to create the 12×12 crossword grid with the 16 3-letter homophones buried in it. However, we did not have such good software for creating crosswords then, and he may have been trying to create the grid by hand — which he found was extremely difficult to do, so the idea was abandoned. A few years later, he came up with his own loosely related idea where letters dropped out of clue answers into the sub-grid, and this was published in 2006 as Dropouts.

After I had created the grid for Co-Star in 2020, I realised that I now had the knowledge and software to create the grid for the homophones crossword, and with a little programming I managed to create some sample grids. My small grid did contain 2 letter “O”s which was a little disappointing as it meant having “OWE” twice in the big grid, however we decided that it wouldn’t affect the overall experience of the crossword so very much.

The computer was meant to try and only have the correct homophone in each 3×3 square, however it seems that computers are not perfect, and a stray VEE was mixing it up in one of the squares. By the time that I realised this, my father had already written most of the clues. He was disinclined to start again, and he commented that it would create a bit of extra interest. This proved to be correct, and maybe it was a fortuitous gremlin in the AI as the V seem to cause more trouble than anything else, without I think being unfair.

I wrote my first crossword clue in this crossword for the answer EYOT. My original proposal was “At last paraglide stealthily onto gunfight island”. This was meant to be a fully Fortnite-themed clue, relating to my own personal journey of when I finally purchased a computer on which I could play Fortnite with my nephew (on the lowest resolution settings, but better than nothing). However, my father was concerned that the crossover between Fortnite players and Listener solvers was likely to be somewhat low, and that the clue wouldn’t make any sense to anyone. But he did accept the compromise solution of skydiving onto a desert island, and there was a desert on the island for a long time in Fortnite, so it was at least still Fortnite-inspired.

There was also a small debate about the title of the crossword, which had the working title “Homophones”. My father did manage to come up with the alternative title “Two-four-one”, which I much preferred as it contained a homophone as well as being thematic.

Thank you to my father for his patience and efforts for both our original attempt in 2000, and our second successful attempt more recently. Thank you also to the editors for their edits to various clues, particularly their revised clue for “NEW DEAL”. Given that it was a crossword about homophones, I really should have thought to check if there were any natural homophones amongst the grid entries, particularly given that there were nude eels swimming around waiting to be caught, so thank you to the editors for picking up on this oversight.