Kaj Sand-Jensen has some advice (Point 10 is my favourite):
- Avoid focus;
- Avoid originality and personality;
- Write long contributions;
- Remove most implications and every speculation;
- Leave out illustrations, particularly good ones;
- Omit necessary steps of reasoning;
- Use many abbreviations and technical terms;
- Suppress humour and flowery language;
- Degrade species and biology to statistical elements; and,
- Quote numerous papers for self-evident statements.
Via Improbable Research, which also links to How to make a scientific lecture unbearable (Yeah, right–I need lots of pointers on that one 
