After our peals of Lessness and Cornwall, the next thing we're going to try with Julia and Nick is Turramurra. It isn't a Pickled Egg method, although I think it featured in the initial discussions about which methods to include; maybe it was even in the "also try" list. It's an easy alternative to Lincolnshire and Yorkshire. I will write more about the method when we have actually rung it, but here are the main features:

• Cambridge place bell order.
• Cambridge above the treble.
• Five dodges in 1-2 across the half lead.
• Three dodges in 3-4 across the half lead.
• Far-dodge-near in 5-6 across the half lead.

Probably it can be reconstructed, or very nearly, from that information.

Having decided on a peal of Turramurra, I started having a look for compositions. This is not so straightforward. Turramurra has BDO falseness, and without investigating the technical details (I confess that I don't have the falseness groups and their consequences at my fingertips), this seems to be quite a lot of falseness and enough to significantly restrict the availability of easy and/or handbell-friendly compositions.

So, this is an opportunity to write about what kind of composition I would like to be able to choose, for a single straightforward Surprise Major method.

You might ask why I want to look for easy compositions, given that I've conducted Horton's Four and Graham John's one-part composition of the Nottingham Eight, which are good achievements even on tower bells. Well, Horton's Four took a long time and a lot of practice, and the Nottingham Eight also required a lot of thorough learning and huge concentration by all the band on the day. For a more "ordinary" peal, especially if I'm not going to ring the tenors, I want a tenors-together composition that's not too complicated to call. The less I have to concentrate on the composition, the more I can concentrate on ringing my bells accurately, which helps with the overall stability. I like compositions with one or both of the following features.

1. A high proportion of coursing for 3-4 or 5-6 or, ideally, both.
2. A multi-part composition with part-ends that I like.

What do I like in a multi-part composition?

• Five-part compositions can be good because the part is fairly short, so there's less to learn. My favourite part end group is the one generated by 13526478, so that the part ends are the lead ends of Plain Bob Minor with the tenors added at the back. I like this because the part ends are easy to check, and the coursing orders are cyclic rotations of the plain course (65324, and so on) which I find easy to remember. The second choice of part end group would be cyclic rotations of rounds on the front six, i.e. 13456278, 14562378 and so on. This makes the part ends easy to check, although the corresponding coursing orders are nothing special. However, five-part compositions tend not to produce large amounts of coursing for 3-4 or 5-6. Usually I would think of five-parts for spliced, rather than for single methods.
• Three-part compositions are good, especially if one of the handbell pairs does the same work in every part. So I would look for either 13425678 and 14235678 as the part ends, or 16342578 and 15346278. It's possible for compositions on these plans to have a reasonable amount of coursing for the fixed pair. Sometimes, three-part compositions with a different part end (12356478 is traditionally popular because it's good for generating CRUs) can be started in a different place so that either 3-4 or 5-6 are fixed - an example is the William Barton composition of Lincolnshire.
• Two-part compositions can sometimes be favourable for handbells. If the part end is 12436578 then 3-4 and 5-6 each do the same work in both halves of the peal - which might also include a reasonable amount of coursing. If the part end is 12563478 or 12654378 then 3-4 and 5-6 swap their work between the two halves of the peal, which could be seen as introducing an element of fairness (either equal easiness or equal pain, depending on the other aspects of the composition).

Sometimes, starting at the snap is good for handbells, if this means that either 3-4 or 5-6 is coursing from the beginning. In Cambridge-above methods, the coursing order at the snap is 52346, with 3-4 coursing. In Bristol, it's 32456, with 5-6 coursing. It's possible to have multi-part compositions with a snap start and finish - for example the Roger Bailey composition of Yorkshire, mentioned in this article about compositions of Rutland.

One-part compositions can be attractive if they have other handbell-friendly features. The main examples of this are the various compositions of Bristol and Cornwall that I have written about, with blocks of five befores that keep both 3-4 and 5-6 coursing for four of the five courses.

Now, back to Turramurra. Looking for compositions in the usual places (ringing.org and complib.org) didn't produce anything attractive. To make sure, I did some computer searching, and there aren't any tenors-together compositions on any of the multi-part plans that I described above. This leaves two options: accept some split tenors, or consider a one-part composition that isn't too complicated.

For the first option, I removed the tenors-together requirement from my computer search, and found some five-part compositions with two courses of split tenors per part. There are several variations, but good examples are this one with three bobs at fifths:

5120 Turramurra Surprise Major
Simon J Gay

V  M  B  H  23456
-----------------
         -  42356
3     -  -  42563
   -     -  35264
-----------------
5 part.

or this one with in/out/fourths:

5120 Turramurra Surprise Major
Simon J Gay

M  B  I/O/4  H  23456
---------------------
             -  42356
   -         -  42563
-       x    -  35264
---------------------
5 part.

These are the same composition, just with the two split tenors courses inserted in a different way. I prefer the first one because the tenor is unaffected. We might ring it if I can either ring the tenors myself or persuade whoever does ring them that the split sections are not too bad. As a point of interest, moving the first home to the end of the part changes the part end to 13456278.

As an alternative, I used CompLib to search for compositions published for other methods that are true to Turramurra. The most interesting result was this composition by Albert Pitman. It's listed with a couple of different starting points, but here I'll show the simplest form.

5056 Turramurra Surprise Major
Albert J Pitman

M  B  W  H  23456
-----------------
      -     52436 |
   -        23564 | A
-        2  34562 |

    3A      62345

    A*      42356

   2     -  25634 | B

    3B      64523

    B*      23456
-----------------
A* = A omitting 2H
B* = B omitting H

This is two five-part compositions joined together. Ringing the A block five times returns to rounds (and has the cyclic part ends mentioned earlier), and ringing the B block five times (starting from rounds) also returns to rounds (with unremarkable part ends).

Joining the two blocks together is achieved by omitting the two homes at the end of the last A block, so that the final coursing order is 52436. This means that the second section of five B blocks would return to 52436, but omitting the final home stops it in the plain course, to come round.

This structure means that the composition can be called without counting the parts - reducing the need to count is always a good strategy. Simply call A blocks until reaching the coursing order 52436, and then call B blocks until returning to the plain course.

I think it's better to ring the A section first, for two reasons. First, because the B section has a before in every course, so the tenors are skipping three leads every time, and I wouldn't want to ring ten (short) courses of an unfamiliar method and then suddenly give the tenors a new pair of place bells. Second, because ringing the A section first gives it the nice part ends of 13456278 and so on.

So I've got a choice of two reasonable compositions, and we can negotiate nearer the time about which one to go for.