SOME SUGGESTIONS ON WRITING A SCIENTIFIC PAPER

Emulate the masters
read the best literature (scientific and non-scientific) to learn style & composition
write for the 21st century, not the 18th
develop an interesting and informative style
use their terminology and style in text, tables, & figures

Write as a scientist
use IUPAC notation
only show significant figures in text
I extend average sig. figs. in data tables to include error sig. figs (e.g., 12.80 + 0.06)
quantify whenever possible (e.g., "large" means nothing)
reference whenever possible, e.g., (Jones, 1997)
be very selective with references (quality journals and recognized experts)
remember the "half-life" of scientific articles is only a few years
only use statistical terms in a statistical sense, and define the statistics, e.g., (P< 0.05, t-test)
delineate assumptions and limitations
eliminate subjectivity
put your work in perspective with current works and ideas (compare & contrast)
be consistent in terminology (e.g., Pb or lead - I find the latter is easier)
acknowledge assumptions, limitations, inconsistencies,(the best defense is a good offense)

Basics
each paragraph should have a topic, which is introduced by a topic sentence
subsequent sentences should substantiate that topic sentence
paragraphs should, usually, have at least 3 sentences
paragraphs should be ordered in a logical sequence, with connective phrases or sentences
maintain same tense when appropriate
use parallel construction in phrases and sentences
use a spell checker, and then carefully proof read
use a grammar checker, and then carefully proof read
have colleagues proof read for both composition and science
include interesting words and ideas
avoid overuse of certain words
minimize use of however, therefore, thus,and try to locate in first or last sentence of a paragraph
data are plural

Suggestions
first spend a few days reading new articles and manuscripts to get a perspective
begin by normalizing and plotting the data to visualize and conceptualize ms.
start ms. with an outline
immediately enter complete citations in "References", including those you may not use rather than delete text, move it to the end of the ms.
also store ideas and points (with numbers and refs.) at the end of the ms.
when completed with the draft, check text, ideas, and points at the end of the ms.
if appropriate, aim for an extrapolation at the end of the ms.
BE DIRECT, BE BRIEF, BE OBJECTIVE