A GUIDE TO NUMBERS, ILLUSTRATIONS AND MEASUREMENTS IN SCIENTIFIC WRITING
Use numerals for numbers of ten or more, e.g.
…the remaining 12 plants were given...
except when the number is the first word in a sentence, e.g.
Twelve sheep were injected with …
Use words for single-digit numbers (less than 10), e.g.
…the remaining five samples were tested for …
except when preceded by a capitalised noun, e.g.
except when part of a sequence of numbers of two or more when one of the numbers is greater
than nine, e.g.
Students were put into groups of 8, 10 and 12.
Treat ordinal numbers like cardinals, e.g.
18 . Dose
Always use numerals with abbreviations and symbols, and when giving a page reference, e.g.
6kg $9 3% 8.3 p.7 p.21
Hyphenate cardinal numbers from 21 to 99 when it is necessary to express them in words, e.g.
Twenty-nine of the repetitions were excluded from the final set.
Do not put two numerals together, as it could cause confusion, e.g.
C:\Docstoc\Working\pdf\85931b27-0e14-4e4e-9bd0-39d1cb9512d2.doc Page 1 of 6
12 15-day-old chickens incorrect
twelve 15-day-old chickens correct
Insert 0 before the decimal point when the value is less than one, e.g.
But do not insert 0 before a decimal when reporting statistics like correlations and levels of
r(38) =29, p< .05
A succession of zeroes may be replaced, e.g.
3.45 x 10 (instead of 34 500 000)
You may also use symbols, e.g.
1MJ (instead of 1 000 000J)
lμj (instead of 0.000 001J)
When specifying a range of numbers, give the numbers in full when they are below 100. Otherwise,
give at least the last two digits, e.g.
2-3 10-12 96-99 103-04 923-29
When the numbers refer to years, use as many numbers as needed to make the time clear, e.g.
In the SI (International System of Units) system, the second (s) is the only
accepted unit. Sub multiples such as the ms are also used.
You may also use min (minute), h (hour), and d (day). Do not use weeks.
Express time in one unit only, e.g. 6.5h, not 6h 3Omin.
Use the four-digit system for clock time, e.g. 1030 EST; 2215 GMT; 0700 CST.
Write calendar dates as 12 March 1990.
Decades may be written in numerals, but with no apostrophe: 1970s.
(This applies to other well-known numbers too, e.g.
He piloted 747s...
Moody made five 50s and three lOOs this summer.)
C:\Docstoc\Working\pdf\85931b27-0e14-4e4e-9bd0-39d1cb9512d2.doc Page 2 of 6
A century should not be capitalised:
this occurred in the seventeenth century
Units and Measurements
Units must always be stated unambiguously. Use the SI units if possible. Choose a unit which is
easily dealt with for your purposes, e.g.
38.2mg rather than 0.0382g or 38 200 μg
Units named after people are spelt without capitals (watt, joule, hertz) although their symbols are
capitals (W, J, Hz).
Do not repeat units in a series of values; e.g.
3 and 4g, not 3g and 4g.
Precision, Accuracy, and Significance
All results should be given with a measure of their statistical variation. You should give the reader
n which means the number of individuals;
)( which means the arithmetic mean ;
s.d. which means the standard deviation;
s.e. which means standard error.
Generally, results should be rounded to a degree no less than 1/10th of the s.d.:
268.92 (s.d. 32321) kg ha-1 &‘ should be written as 270 (s.d. 323)kg ha-1 d-’.
The standard deviation should always have one more decimal place than the
mean to which it refers.
Use the most manageable unit (see above, UNITS), except that a dimensionless fraction with a
maximum value of one should be left in that form or expressed as a percentage.
Pasture DM yields and DM growth rates are always in kg ha-1 cF1.
Do not quote a result to more significance than its accuracy justifies.
This means that if you say “5.OOg’ it should have been measured on an analytical balance,
and someone replicating your experiment would need to do the same. If you used a
rougher measure, then only say “5g”.
Express concentrations of substance as mass or moles per unit not as percentages, as this
could cause confusion when the experiment is replicated.
Diagrams and illustrations
Illustrations, in the form of tables, graphs, diagrams, maps, photos, line drawings, and charts, are
often an advantage in a scientific paper. However, they are not used merely to improve the
C:\Docstoc\Working\pdf\85931b27-0e14-4e4e-9bd0-39d1cb9512d2.doc Page 3 of 6
presentation of the paper, nor to take up space when wordage is short They may be used only as a
visual aid which will express features otherwise difficult to grasp in the text.
Any such illustration should be placed as close as possible to its first mention in the text. It should
not duplicate information given elsewhere in the paper.
Tables are used for the presentation of extensive numerical results. They are not restatements of
information given already, and should not be used to show mailer which can be easily stated in the
Tables are usually numbered with an arabic numeral (e.g. Table 3) and captioned above. The
source and any notes to do with the table are given immediately below it.
The caption is a brief, self-explanatory statement of the nature of the facts in the table. The caption
is not usually a sentence; it is not necessary to say ‘a summary of’ or ‘a statement of’.
State the units for every quantity in the table, but do not put units in the caption. Units should be
placed at the head of each column or group of columns. Record measurements of the same
quantity vertically in columns.
Do not give numbers in tables too many significant digits. If the numbers in a table represent the
mean of several values, always give some measure of the variation involved, preferably the
standard error of a difference or mean.
Where data have been transformed for statistical analysis, give the original means in the table
rather than the transformed. However, standard errors appear only in their transformed state.
SEE THE EXAMPLES OF TABLES WHICH FOLLOW
Table 1. Effect of glasshouse and field environments on seed weight of five cultivars of nasturtium.
Seedweight (g) Seedweight (g)
Red Star 4.6 3.3
Rosa Red 3.2 2.9
Nancy Red 3.1 3.7
Red Dragon 4.2 3.9
Double Red 4.9 3.1
C:\Docstoc\Working\pdf\85931b27-0e14-4e4e-9bd0-39d1cb9512d2.doc Page 4 of 6
Table 2. Primary School Enrolments by area over five.years.
Year Barwon Headland River St Church Hill
1992 1,027 932 1,576 2,432
1993 1,211 786 1,833 2,680
1994 1,292 911 1,984 2,120
1995 * 2,853 1,378 1,730 1,843
1996 3,178 1,932 1,420 1,572
Graphs and Other Illustrations
Graphs can be used as a visual aid to help express features which would be less obvious in a
Graphs can neatly summarise matter which would be too long in a table, and can show an
empirical relationship between two quantities.
Graphs are called ‘Figures’. in a paper, and are given an arabic numeral, with caption, label, or
title, just below. Such ‘Figures’ are always mentioned in the text with a capital, and may be
abbreviated as in ‘Fig. 1’.
Any Figure (photo, map, drawing, graph, chart) should be shown with a clear margin all around. A
full-page Figure is numbered consecutively with the text (not, for example, as ‘p.4a’), and has a
margin of 2.5cm all around; a greater margin should be given on the left-hand side if the paper is to
be bound. Everything relating to the Figure should be within this margin. If the Figure is very wide,
it may be set broadside in the paper, in which case the caption will be broadside as well. The key
or scale for the Figure should be placed within or beside the Figure, rather than below it.
Underline identifying information in the tags before the item being identified, e.g. ‘Above left, the
sheep prior to testing; below left, after testing
Rules for Graphs
Axes: Make sure these are graduated at sensible intervals, so that the experimental points are not
all cramped together by the use of too small a scale. Number every second interval, to avoid
confusion from too many numerals.
C:\Docstoc\Working\pdf\85931b27-0e14-4e4e-9bd0-39d1cb9512d2.doc Page 5 of 6
Represent the independent variable (Le., the one decided by the experimenter) on the horizontal
axis (x-axis or abscissa). Use the vertical axis (y-axis or ordinate) to show the corresponding
Label each axis with the name of its variable and its corresponding units, either as, for example,
Nitrogen/kg ha- or ‘Nitrogen (kg ha4)’.
Try to keep variable names short, and not to label the axis with numbers of more than three digits;
instead of ‘2000kg ha write ‘2t ha-1’.
You may use full (_), pecked (---), or dotted (. .) lines to indicate the experimental results. Use
distinctive symbols ( A 0 I]), or filled symbols (A S ), to indicate experimental points.
Graphs should show a trend in a smooth curve or straight line, instead of bending erratically to
embrace isolated points. To help clarity, try to use no more than four lines on any graph.
C:\Docstoc\Working\pdf\85931b27-0e14-4e4e-9bd0-39d1cb9512d2.doc Page 6 of 6