"Goodness!" he ejaculated.

Back in olden times when I was at school, one of our favourite sources of snigger moments were the pages in Lord Baden-Powell's autobiography where instead of using 'I said' he uses 'I ejaculated'. We may have been young, we may have been silly, but we knew for sure that using the word ejaculated to refer to speech was not a good idea. (And especially when boy scouts are involved.)

So imagine my surprise to come across a How To Write book in the library recently that seriously suggested 100 alternatives to 'said', including the e word. Most of them were simply wrong. Take 'Woody Allen is so funny,' she laughed. Okay, now say it aloud, at the same time as laughing. Try it. Then try snorting your words, or giggling them. It can't be done. You can laugh, or snort, or giggle, and then speak, but not both actions at the same time.

I'm all for adding colour to writing, but in the case of speech the only variations that work are about volume: 'Be quiet,' he whispered, or 'Stop that now,' she shouted. In general, plain old said is safest. It's a word that disappears, like 'and' or 'the'. If you feel you've got too many 'saids' floating around in your writing then either look for ways to hide them (eg within speeches), or cut them completely (substituting actions if necessary). 'But don't be tempted to use highly coloured alternatives for said,' she ejaculated. Unless you want us to snigger.

Writing without Chapters

A chapter is a useful tool for the reader. It divides the novel up into easily manageable sections so the reader can spread out the contents over several days or weeks, perhaps a chapter before bedtime.

A chapter is a useful tool for the writer. It divides the novel up into easily manageable sections so the writer can spread the labour of writing the darn thing. It makes it easy to plan a book - say, three scenes per chapter of about 1500 -2000 words each scene, and twenty scenes - and there you are. Novel written.

Except it's not that easy. A chapter is not a useful tool for good story telling. A chapter is not a useful tool for rewriting. A chapter is not a useful tool for rearranging. Okay, I'm going to go headlong against those who like to plan out their novel before they start writing, but in my opinion a chapter is not a useful tool for writing a novel that works.

Writing by chapters inhibits creativity by arranging it into nice chunks. It's the Tick Box approach to writing, no deviations allowed. I've heard writers say that they couldn't possibly move this scene some place else, even though they can see why it's been suggested, because then the chapter would be too short. And rewriting is often out because it upsets chapter balance. And the amazing cliff-hanger which will have the readers turning the pages faster than a Zeotrope machine can't possibly go there because it is ordained that the chapter finishes six pages later on.

Sectioning the novel into chapters is about the last thing I do before it goes off to my editor. They may be between 1000-6000 words, but I'm looking for variety in length and brilliant chapter ends. As the novel gets towards the end, the chapters become shorter to help pick up the pace. Above all, the chapters go where it suits the story-telling and not the other way around.

Which books would you save from the flood?

According to www.thisiswales.co.uk when Waterstones Swansea was flooded they saved the books on Hitler. It made me wonder what I'd save out of the 10,000 or so books I have about the house. I think I'm going to let all the reference books go, even though it's taken me years to build up a decent collection. Of other non-fiction, the only book I'm keeping is Antonia Fraser's wonderful study of women in the C17th because it made such an impression on me when I first read it, and I hope I'd be inspired to rebuild my post-flood life by the stories of women such as Lady Brilliana Harvey.

On to fiction. There are books that I have loved desperately and passionately but - rather like former boyfriends - I haven't thought of in years and suspect they won't stand up to viewing today. Authors like Jean Plaidy and - gulp - Georgette Heyer, although I will have to save Venetia. Sadly, they will have to take their chances with the water, the exception being Elizabeth Goudge and The Dean's Watch, The Little White Horse and the Damerosehay trilogy. Of the classics a certain degree of literary honesty will be required, and Camus, Tolstoy and Woolf will be left behind, along with Dickens, who I've never really enjoyed. Austen will definitely be coming with me, and if there's only room for one, it'll be Persuasion. However, I'm hoping there's time to grab the complete works and Jon Spence's Becoming Jane, which I preferred to Claire Tomalin's biography.

Moving on to more recent reads, Anne Tyler is essential, though if space is limited it'll only be the books with a male protagonist as I find her women protagonists a bit wet and in the circumstances that's unfortunate. William Trevor's Complete Short Stories will sustain me with their gentle wit and intelligence as will Nadine Gordimer's. My groaning To Be Read shelf will have to be left behind, so I'll miss out on Steig Larsson who was next up.

And if there's only one book to save? Shakespeare's Complete Sonnets. The wisdom of the world in one pocket sized volume. Which books would you choose?

Chapter Ends

I'll let you into a secret: I'm a bit obsessed by chapter ends. It's one of those little corners that seem to be left out of books on writing, but it's an important area for a writer to master. Think about it. Most people read late at night. They want to read a chapter before turning the light out. Your job is to get them to read another chapter, and another. And another, all the way until The End.

Watch soaps for masterclasses in how to get people tuning in the next day. There's always a sudden revelation, a question that must be answered or a dramatic situation to be resolved. When writing we might choose to be more subtle about it, but essentially the trick is the same. Make the reader read 'just a little bit' of the next chapter, and you've got them hooked.

Some writers do this naturally. JK Rowling is a good example, as I discovered when I read the first Harry Potter books aloud to my children. The chapters are long and it's hard to find a natural point at which to stop. Unless you want to read for an hour you end up breaking in the middle of paragraphs. Small wonder she gets kids reading; the books are compulsive page turners because there are no places to stop.

The worst thing is to end the chapter with them all going to bed and zzzzz-ing away - you might as well write 'put this book down now'. The exception is when you're writing a picture book, when parents are reading hoping their children will drop off at the end of it. So, if you want to write a page turner, pay attention to your chapter ends. Keep them reading.

How to choose an agent

How to choose an agent is the sort of statement that would have had me rolling my eyes in disgust before I acquired one because there's definitely a point when a would be writer feels that ANY agent will do. But any agent won't do, it's got to be the right one for you. Some things to think about...

1 Do you want an agent who does a lot of editing on your work before sending it out to publishers? Most agents will have worked at publishing companies before going over to the dark side. A former editor may be unable to resist the urge to edit your work which, depending on your point of view, may be a good thing. On the other hand, you might prefer the sort of agent who settles down happily with a stack of miniscule-print contracts and enjoys quibbling over percentage points, in which case a background in rights would be good.

2 Do you want to be a big fish in a little pond? In which case you want a small agency. Your agent's income will be directly linked to yours, so they've got extra reason to sell your books (and generally be nice to you!). A big agency may make you feel they're too busy dealing with their star authors to have time for you. The number of authors each agent represents is also relevant - see also 6.

3 But a small agency may not have as much clout as a large agency, nor are they likely to have the same range of experience ie they tend to specialise in one kind of writing. A larger agency however will have specialist departments in children's writing, television/film etc and may have offices around the world. (Smaller agencies will have agreements with other agencies to bring in the expertise when needed, so this shouldn't be a dealbreaker.)

4 Personality. Top of my wish list was the desire not to be frightened by my agent and some literary agents I have met are terrifying. But I'm easily scared. Some writer friends relish the formidable qualities of their agent. Others couldn't care less so long as the agent does their job.

5 Age. Young agents are enthusiastic and keen to make their mark, but they may not have either the contacts in publishing or the experience. On the other hand, they may be actively looking for clients. Older agents have shedloads of experience and contacts - and existing clients.

6 Money. No reputable agent should ask you for money up front. Full stop. As to the percentage, I'd prefer they took 15% and represented fewer authors. But you might be happier with 10%.

Think about what's going to suit you. Go to talks at literary festivals and conferences, ask other writers, join writing societies and groups and generally get out there and do some research before drawing up your 'hit list'. Only an idiot would go to all the expense and effort of getting a positive response from an agent and then go all wobbly at the thought of being represented by them. Don't be that idiot.

Does Size Matter?

Size is something that writers tend not to talk about except in reference to other people, as in 'Did you hear how big it was?' Of course you want to imply that you've got a whopper (while being suitably modest about your attributes) because size is linked to desirability, but you don't want to actually tell anybody your vital statistics unless it's so large it hits the headlines, in which case everybody knows. But does size really matter?

I am, of course, talking about advances. Advances are what the publisher gives a writer on signing them up. It's not magic money, it's an advance against the royalties that the publisher is guessing you might make from the book. Until you've earned out your advance - ie the royalties you would have received exceed the money you've already had - you won't see a penny more cash. Publishing wisdom dictates that the amount paid out in the advance is roughly equivalent to what they'll spend on marketing: expensive books get big marketing campaigns. So you'd think that a big advance was good news (and if my publisher is reading this, they undoubtedly are) but they can be a double edged sword for a new writer.

Let's say you've got a six figure advance for a two book deal. This usually will be for the wonderful book you've written and a synopsis or outline of the fabulous book you're hoping to write next. You're thrilled, your agent is thrilled, the publisher is thrilled. You get on with writing Book No 2 while the publisher spends a lot of money on the marketing. Expectations are high. Book No 1 comes out and does quite well but you'd have to sell an awful lot of books to cover your advance and you don't sell that many. The publisher loses money. You hand over Book No 2 which you think is even better than Book No 1, but no one is thrilled any more. Book No 2 comes out with no marketing so doesn't sell (the publisher may even decide to cut their losses and not print it). Your name is forever tarred as the author who lost the publisher lots of money.

Contrast the situation with the writer who got a modest advance so Book No 1 doesn't have to sell that many copies before it starts earning out. It achieves those sales. Everybody is happy and looks forward to Book No 2, which also earns out, with sales a little bit more than Book No 1. The sales graph is on an upward curve, the author is a success and gets offered a new contract.

Over the last ten years I've met three writers who fell into the first category. None of them had got beyond Book No 2 and the advance money was long gone. They felt cheated. Bitter, even. All writing is like a high wire act, but if you have a big advance there's further to fall. Rumour has it that, with the recession, advances have got smaller and maybe that's a good thing for writers who want long term careers. Still, I can't see any writer turning down a whopping big advance. I certainly wouldn't.

Flashback - Work of the Devil

I am on a one-woman campaign to eliminate flashback from writing. I don't understand the when it usually makes writing soggy and dull. Don't do it! Writing is partly about controlling pace. Action and dialogue speed it up, reflection and description slow it down. Flashback stops the novel dead in its tracks, like forcing a speeding car into reverse. Why? Because it's already happened. When you read you want to know what's going to happen next. With a flashback you know. Think how many stories have an underlying Will they, won't they? premise. If you use flashback, you've lost it.

I recently workshopped a student piece. It opened with a scene of the main character on her way somewhere, then went into flashback for the rest of the chapter going over the previous 24 hours. It was well written, but the tension was lost - we already knew the character was going to get in the car and head off for her destination, because the writer had said so at the beginning. It's like telling the punchline before you do the rest of the joke. It doesn't work.

Worse are those flashbacks where the character settles down to have a jolly good reminiscence like some great-uncle at Christmas beginning with 'I remember those lovely holidays we used to have at Gran's...' For some reason, they're often inspired by train journeys and as the character drops off into a dream of times past, so do we.

Of course there are exceptions. Donna Tartt's The Secret History gains a lot of tension because we know from the start that a main character is going to be murdered by one of the others. But we don't know who murders whom until the end. Anita Shreeve's The Pilot's Wife uses flashback to unravel the story, but each flashback section advances the plot with information the reader didn't know until that moment. So flashback can work, but the chances are more likely that they're losing tension and pace. Don't do it!