A|B|C|D|E|F|G|H|I|J|K|L|M|N|O|P|Q|R|S|T|U|V|W|X|Y|Z|ALL

  检 索： 期刊名称 ISSN         高级检索

期刊名称：JOURNAL OF THE SCIENCE OF FOOD AND AGRICULTURE ISSN： 0022-5142 版本： SCI-CDE 出版频率： Monthly 出版社： WILEY, 111 RIVER ST, HOBOKEN, USA, NJ, 07030-5774   出版社网址： http://as.wiley.com/WileyCDA/ 期刊网址： http://as.wiley.com/WileyCDA/WileyTitle/productCd-JSFA.html 影响因子： 3.639 主题范畴： AGRICULTURE, MULTIDISCIPLINARY;    CHEMISTRY, APPLIED;    FOOD SCIENCE & TECHNOLOGY 期刊简介(About the journal)    投稿须知(Instructions to Authors)    编辑部信息(Editorial Board)    About the journal The Journal of the Science of Food and Agriculture publishes peer-reviewed original research, reviews, mini-reviews, perspectives and spotlights in these areas, with particular emphasis on interdisciplinary studies at the agriculture/ food interface. This international journal covers fundamental and applied research, including: Food - Health and Nutrition: Human and animal nutrition. Enhancing health and wellness with novel and functional foods, nutraceuticals (but not pharmaceuticals) or food supplementation. Evaluation, including chemical tests, animal models and in vitro, in vivo testing systems. Food Science and Technology: Development, production and processing of human food, animal food, and beverages. Measurement of food and ingredient characteristics and functionalities. Papers reporting simply antioxidant activity and those reporting the composition of foods that are not widely eaten outside a small geographical area will not generally be considered for publication. Food Safety: Microbiological food safety, including on-farm safety management, epidemiology of food-borne illness, novel process and packaging, and detection of food-borne pathogens. Food toxicology and safety evaluation of genetically modified organisms. Application of risk assessment and risk management techniques. Biotechnology: Applications of biotechnology to plants, animals and microorganisms as they relate to agriculture, food production and processing, and fermentation technology. Molecular Biology: Applications of genetic modification and molecular biology, bioinformatics and ?omics? technology to plants, animals and microorganisms as they relate to agriculture, food production and processing, and fermentation technology. Biochemistry: Novel insights into the chemistry and biochemistry of foods, feeds, ingredients and components, including macro-, micro- and non-nutrients, and also enzymes and other processing aids and functional ingredients. Research that addresses current issues within the agri-food area, for example functional foods in relation to health claims and safety. Materials and Processing: Food raw materials and quality. Food structure and physical properties. Physical description of processing and product properties of raw materials. Food process engineering. Sensory and Consumer Sciences: Perception and evaluation of characteristics of foods and beverages including flavor, appearance and texture. Consumer behavior issues, including acceptability, preference and other factors related to cultural cues and product image. Agriculture - Production: Plant and animal husbandry including precision agriculture and aquaculture and mariculture. Impact of scientific and technological developments on yields and quality. Agriculture - Utilization: Plants, plant products, animals, animal products, fish and marine products as human food or animal feed. Industrial applications for plants, plant products, animal products and marine products. Agriculture - Environment: The interaction of agriculture and environment, including safety and sustainability of agricultural practices; protection of natural resources (quality of air, water and soil); mitigation of environmental effects through improved farming methods; assessment of organic farming and alternative production systems; management of nutrients, pathogens or other potential contaminants; impacts and mitigation of climate change. Biomass and Bioenergy: Biomass production and utilization, feedstocks, biodegradation methodologies, potential products. Many studies are likely to overlap several subject areas. Manuscripts will be allocated to the editor whose primary interests are appropriate to the main subject of the paper. Instructions to Authors GENERAL Journal of the Science of Food and Agriculture publishes original research and critical reviews in these areas, with particular emphasis on interdisciplinary studies at the agriculture/food interface. The Journal covers fundamental and applied research in many areas including: health and nutrition of food food science and technology biotechnology molecular biology biochemistry of foods. feeds, ingredients and components food safety materials and processing agricultural production agricultural utilisation agriculture and the environment sensory and consumer sciences. Please read the Aims and Scope section at the front of the journal for a more detailed view. Concise contributions on experimental or theoretical investigations and critical reviews of the science of food and agriculture are invited for publication. SUBMISSION OF PAPERS There are two options: 1) JSFA operates an on-line submission system. Details of how to submit online and full author instructions can be found at: http://sci-wiley.manuscriptcentral.com/. Referees may ask to see hard copies of electronic figures for clarification; these must be available immediatley on request. 2) Alternatively, send one original and three copies of the printed manuscript, with any tables or illustrations and an electronic copy and covering leter, to one of the receiving centres below (all hard-copy submissions will be entered into the online database and may be tracked by the author online. An explanatory e-mail will be sent to the corresponding author). Receiving Centres: The Journals Manager Journal of the Science of Food and Agriculture SCI, 14/15 Belgrave Square London SW1X 8PS UK Tel: +44 (0)20 7235 3681 Fax: +44 (0)20 7235 0887 e-mail: jsfa@soci.org Authors in North America should submit to: Professor D S Reid Department of Food Science & Technology University of California, Davis 1 Shields Avenue Davis, CA 95616-8571 USA Tel: +1 530 752 8448 Fax: +1 530 752 4759 e-mail: dsreid@ucdavis.edu All papers must be in English. Non-English speaking authors who do not have a good command of English are advised to seek assistance from someone who has. Authors in Japan, please note: Wiley-Japan can provide authors in Japan with a list of recommended services to check and improve the English in papers before submission. Please contact Masayo Kobayashi in the Wiley-Japan office by fax (+81 (0)3 3556 9763) or e-mail (editorial@wiley.co.jp) for more information, stating which Wiley journal you wish to submit to. Inadequately or incorrectly prepared typescripts may be rejected. Authors will receive an immediate acknowledgement of receipt of their paper followed, normally within four months, by a decision. When preparing a manuscript, authors should refer to a recent (2003 onwards) issue of the Journal and follow the detailed instructions given below. The corresponding author must obtain the consent of all the co-authors to the submission of the paper. Papers will only be accepted on the understanding that their contents have neither been published nor are being offered for publication elsewhere. In order to expedite the manuscript review process, authors can indicate the subject area they believe to be most appropriate to the topic of their manuscript (from the subject headings listed above, which are described in more detail in the Aims and Scope). MANUSCRIPT PREPARATION Length Papers should not normally exceed 6000 words, including relevant data. Any manuscript that is submitted which, in the opinion of the editor, is too long will be returned to the corresponding author for redrafting within a suggested maximum wordage. Typescript Type papers in double spacing on one side of A4 or 8½" ¡Á 11" paper with 30 mm wide left and right margins. Underlining to indicate italicised type (or use of italic type in the manuscript) should be restricted to genera and species names, chemical descriptors (eg cis, trans etc) and journal and book titles. Do not underline any headings. Footnotes should be kept to a minimum and indicated by * or †. Do not use full stops after abbreviations unless essential for clarity. Abbreviations of chemical and other names should be defined when first mentioned in the body of the paper, unless commonly used and internationally known and accepted. Each page should be numbered individually. Units and nomenclature Units Use SI units in accord with the recommendations of the International Organisation for Standardisation (ISO). Use the form g kg-1 etc (not %) to specify content/composition/concentration. Use % only to express proportional change. Note that the form g 100g-1 etc is not correct. Avoid the use of g per 100g, for example in food/feed composition, by using g kg-1. Fertiliser rates should be presented in terms of the element applied. Further information on the ISO recommendations can be obtained from the following publication issued by the British Standards Institution, London: Specification for SI units and recommendations for the use of their multiples and of certain other units, BS 5555 : 1993 ISO 1000 : 1992. Symbols Write all symbols, formulae and equations with great care. Unusual symbols (including Greek lettering) should be defined in words in the left margin at the first mention. Scientific names Give the scientific names (with authority) for plants, animals, microorganisms, with generic names in full at the first mention, eg Myzus persicae (Sulzer). Thereafter abbreviate them in the text, eg M persicae; give them in full (without authority) in the headings of sections, tables, figures and key words. Where appropriate, cultivars should be specified. Enzyme nomenclature Identify each enzyme together with its EC number, if available, at the first mention, following the recommendations of the latest edition of Enzyme Nomenclature. Chemical nomenclature Use the current systematic IUPAC nomenclature throughout. Statistical analyses Particular care should be taken to ensure that the appropriate statistical analyses have been carried out. The methods used should be described concisely, yet with enough information to explain how the chosen methods have been applied to the data. The form of all experimental errors and their statistical significance must be given clearly. The statistical analyses should be used in the discussion to justify inferences made against the background of normal biological variation. Further information on recommended statistical procedures can be found in Appendix II, or printed in J Sci Food Agric 81 No 1 (2001); additional copies are available from the Managing Editor. Layout The main body of the paper should be divided into unnumbered sections and each given an appropriate heading. Main headings should be capitalised and centred over the text. Choice of headings will depend on the content, but the following is recommended for research papers: Title This should be concise, specific and explain the nature of the work. State in a footnote if the paper was given, in whole or in part, at a scientific meeting. Running title A running title of up to 80 characters should also be provided. Authors' names Each must have the customary forename in full and initials (eg Arthur B Smith). Give the full address(es) where the work was done. If an author was on secondment or visiting from another address, or has since moved to a new address, this should be given in a footnote. Abstract This must be informative yet concise, give essential information such as the purpose of the work, the data derived from it and their statistical significance, and be intelligible without reference to the paper itself. It should not normally exceed 150-200 words. Authors should remember that the abstract is often the only portion of a paper read, as in abstracting journals, and the use of unusual acronyms or abbreviations should be avoided. Key words List all the main topics incorporated in the paper, including any already given in the title. Introduction Include a clear description of the aims of the investigation (without summarising the work itself) and a brief statement of previous relevant work with references. Experimental State clearly, in sufficient detail to permit the work to be repeated, the methods and materials used. Only new techniques and modifications to known methods need to be described in detail but known methods must have adequate references. Include the name, postal town, code and country of the supplier or manufacturer of any chemical or apparatus not in common use. Give the statistical design (including replication) of each experiment where appropriate (see also Statistical analyses, above). Results Present these concisely, using tables or illustrations for clarity; do not list the results again in the text. State clearly the form of the experimental error and the statistical significance of the results (see also Statistical analyses, above). Do not overstate the precision of the measurements. Histograms or bar charts, unless prepared carefully, are inferior to tables. Only in exceptional circumstances will both tables and illustrations based on them be accepted. The Experimental and Results sections may be combined when appropriate. Discussion The Results should be followed by a concise section to discuss and interpret them. Do not just repeat the results. A combined Results and Discussion section sometimes simplifies the presentation. Conclusions Do not merely repeat content of preceding sections. The Discussion and Conclusions sections may be merged. Acknowledgements Keep these to the absolute minimum. Avoid thanks for permission to publish. References Check carefully for accuracy and follow the correct style. Refer to unpublished work only in the text (Smith A B unpublished), (Brown C D pers comm). Indicate literature references at the appropriate place in the text using superscript numbers in the order in which they appear and a full numerical list must appear at the end of the paper, giving all authors with initials after the respective surname. Ensure that all references in the list are cited in the text and vice versa. Give the date and full title of the paper in the language in which it appeared or an accurate English translation. Abbreviate all journal titles as in Chemical Abstracts or Biological Abstracts and the annual BIOSIS List of Serials, without using full stops after abbreviation. If the journal is not included, give its title in full. Volume numbers should be bold. Note the following style and order for journals: Syers JK, Mackay AD, Brown MW and Currie LD, Chemical and physical characteristics of phosphate rock materials of varying reactivity. J Sci Food Agric 37:1057-1064 (1986). Quote books as follows: Doyle J, Altered Harvest. Agriculture, Genetics and the Fate of the World's Food Supply. Viking Penguin Inc, New York, pp 136-158 (1985). Thomas T, Barnes A and Hole CC, Modification of plant part relationships in vegetable crops, in Chemical Manipulation of Growth and Development, Ed by McLaren JS. Butterworths, London, pp 297-311 (1982). When quoting patents give the name of the applicant, the year of publication, the title, the country and patent or application number thus: Hagner MB and Wendt KL, Method of sorting seeds. UK Patent 1470133 (1977). Tables Supply each table on a separate sheet. The table number (given as an Arabic numeral) should be given at the top, followed by a concise title. Give essential details as footnotes. Keep the number of columns to a minimum. Column headings should be brief, with the units of measurement clearly stated in parentheses. Where one unit applies to all the data in the body of the table include it in the title. The data should be easy to follow without horizontal lines between entries. A zero is often incorrect; use 'not detected' (ND) where appropriate, amplifying this, and trace (tr), where possible in a footnote, eg '¡Ü10 µg kg-1'. For 'not significant' (NS) state the limiting level. Illustrations Include only if essential, and number line drawings, figures and photographs in a single sequence in order of appearance using Arabic numerals. Keep lettering and numbering (characters) on illustrations to a minimum and include essential details in the legend. Write the authors' names and the title of the paper on the back of each illustration using soft pencil. Where any confusion is possible identify the top of the illustration clearly. Photomicrographs must have a scale bar. All legends should be combined in one list and typed on a separate sheet. Where possible, illustrations should also be submitted in electronic format. Save each figure as a separate file, in TIFF or EPS format preferably, and include the source file. Write on the disk the software used to create the files. Use dedicated illustration packages in preference to tools such as Excel or Powerpoint. Line drawings and figures should be in a form suitable for direct reproduction, no larger than A4 or 8½" ¡Á 11", in black ink, with stencilled lettering (avoid using dry transfer, typewritten or handwritten lettering) all in proportion to the amount of detail. Computer-drawn diagrams must be prepared on a high quality laser or ink jet printer or plotter, not on a dot matrix printer or equivalent. Use only essential characters and insert these and any other symbols clearly; explain all symbols used, and where a key to symbols is required, please include this in the artwork itself, not in the figure legend. On graphs, include labels and units on axes. Present logarithmic scales with arithmetic numbering 0.1, 1, 10, 100 rather than -1, 0, 1, 2. Avoid unnecessarily long axes that lead to large blank spaces on graphs. Line drawings and figures should all require the same degree of reduction and all characters must be chosen so that after reduction they are at least 1.5 mm in height. The type area of the journal is 172 mm wide ¡Á 249 mm deep, in two columns each 81 mm wide, and the characters should therefore be large enough to be legible after reduction of the illustrations to fit the page or column width. Photographs (halftones) should be supplied as glossy prints (four original prints of each) of good contrast, photocopies are not acceptable. Do not allow them to be damaged by paper clips, folding etc. Some loss of clarity may occur during reproduction. Electrophoresis patterns These are complex. Photographs, which often lack clarity, should not be included except to make a particular point. Where the reporting of gel electrophoresis, SDS gels, immuno-electrophoresis, isoelectric focusing etc is essential, adhere to the following principles: a single zone requires only description in the text preferably claim homogeneity using a scan diagram preferably use a single gel to compare several tracks when scan diagrams are used accurate alignment is essential Where photographs or scan diagrams must be used: number all zones and identify those common to more than one track give a molecular weight scale for SDS gels give experimental details and track identification in the legend Please note: the journal does accept colour figures and photographs, but there is a charge to authors to cover the additional production costs involved in printing colour. Under exceptional circumstances, authors may request that these charges be waived. This must be provided in writing, at the time of submission, and authors must justify to the Editor that inclusion of the figure(s) in colour is essential for interpretation of the results presented. Chemical structures Prepare these on a separate sheet as described for illustrations and number the individual formulae with Roman numerals (I, II). All bonds, charges and free radicals should be accurately positioned. Indicate aromatic and unsaturated heterocyclic systems using double bonds. Preferably use general structures, distinguishing related compounds by substituents R1, R2 etc. PERMISSION GRANTS If the manuscript contains extracts, including illustrations from other copyright works (including material from on-line or internet sources), it is the author's responsibility to obtain written permission from the owners of the publishing rights to reproduce such extracts using the Wiley Permission request Form. Permision grants should be submitted with the manuscript. CRITICAL REVIEWS Authors wishing to submit a critical review should first submit an outline proposal to the Journal Manager. BIOETHICS Author submitting a manuscript containing in vivo animal work should submit details of all relevant Ethics Comittee approval and authorization (eg institute and/or government) and all reference numbers. Details will be printed as a footnote to the paper. PROOFS The address to which the author's proofs are to be sent must be indicated clearly. Proofs will be despatched to the corresponding author. we will email pdf proofs where possible. These must be corrected and returned to the publishers within 48 hours of receipt; failure to do this will result in a delay to publication. Author's corrections must be restricted to printer's and/or factual errors. OFFPRINTS There are no page charges. Twenty-five offprints of each article are supplied free of charge. Additional offprints can be ordered at current printing prices. COPYRIGHT Upon acceptance of a paper by the Journal, the author(s) will be asked to sign a transfer of copyright of the paper to SCI. The transfer will ensure the widest possible dissemination of information. Where an article is prepared jointly, the corresponding author must obtain either the signature(s) of the co-author(s) to this Agreement or their written behalf. APPENDIX I: QUANTITIES, UNITS AND SYMBOLS Recommended SI units and symbols SI base units Name Symbol Length metre m Mass kilogram kg Time second s Electric current ampere A Thermodynamic temperature kelvin K Amount of substance mole mol Luminous intensity candela cd Multiples To form decimal multiples of SI units the following prefixes may be used but for mass the prefix is added to the gram (g) and not the kilogram (kg). Multiple Prefix Symbol 103 kilo k 10 deca da 10-1 deci d 10-2 centi c 10-3 milli m Multiple Prefix Symbol 10-6 micro µ 10-9 nano n 10-12 pico p 10-15 femto f Derived units Physical quantity Name Symbol Definition Energy joule J kg m2 s-2 = Nm Force newton N kg m s-2 = J m-1 Pressure pascal Pa kg m-1 s-2 = N m-2 Power watt W kg m2 s-3 = J s-1 Electrical charge coulomb C A s Electrical potential difference volt V kg m2 s-3 A-1 = J A-1 s-1 Electrical resistance ohm kg m2 s-3 A-2 = V A-1 Electrical conductance siemens S kg-1 m-2 A2 = -1 Electrical capacitance farad F A2 s4 kg-1 m-2 = A s V-1 Magnetic flux weber Wb kg m2 s-2 A-1 = V s Inductance henry H kg m2 s-2 A-2 = V s A-1 Magnetic flux density tesla T kg s-2 A-1 = V s m-2 Luminous flux lumen lm cd sr Illuminance lux lx cd sr m-2 Frequency hertz Hz cycle s-1 = s-1 Customary temperature degree celsius ¡ãC K Mass tonne t 103 kg = Mg Area hectare ha 103 m2 Volume litre litre 10-3 m3 = dm3 Other derived units Physical quantity SI unit Symbol Area square metre m2 Volume cubic metre m3 Density kilogram per cubic metre kg m-3 Velocity metre per second m s-1 Angular velocity radian per second rad s-1 Acceleration metre per second squared m s-2 Kinematic viscosity square metre per second m2 s-1 Dynamic viscosity newton second per square metre N s m-2 Electric field strength volt per metre V m-1 Magnetic field strength ampere per metre A m-1 Luminance candela per square metre cd m-2 Entropy joule per kelvin J K-1 Specific heat joule per kilogram kelvin J kg-1 K-1 Radiant intensity watt per steradian W sr-1 Molality mole per kilogram mol kg-1 Amount-of-substance concentration mole per cubic metre mol m-3 Stress newton per square metre N m-2 Young's modulus newton per square metre N m-2 Bulk modulus newton per square metre N m-2 Compressibility square metre per newton m2 N-1 Moment of inertia kilogram square metre kg m2 APPENDIX II: Recommendations of the editorial board on use of statistics in papers submitted to JSFA Guidelines based on a report by A W A Murray (J Sci Food Agric vol 42 (1988)) and updated by J W McNicol* (October 1998) *BioSS, SCRI, Invergowrie, Dundee, DD2 5DA, UK. INTRODUCTION This document is intended to provide some guidance on standards of statistical methodology and presentation for papers to be acceptable for publication in Journal of the Science of Food and Agriculture. It is not a definitive statistical text nor are any rules intended to be inflexible. Good statistical practice, well presented in a manner appropriate to the journal style, should always be acceptable, but certain practices, which may either mislead or not do full justice to the authors' data, will be actively discouraged. These notes comprise some explanatory text and a list of useful references. There are two somewhat interlinked aspects of the use of statistics in the context of scientific papers. These are methodology and presentation. METHODOLOGY The statistical methods used by authors should take account of the following points. Experimental design Replication of treatments The statistical methods used in biology should assist in separating effects of interest from the 'noise' due to the inherent variability of biological systems, and also enable the estimation of quantities of interest together with appropriate measures of their precision of estimation. Authors must take account of the variability of their material when designing experiments. Adequate replication of treatments or samples must be made so that worthwhile inferences may be drawn in the presence of random variation. It is worthless to report 'no significant differences', or effects of similar size to error, when the replication is not adequate to detect differences of a size which would be of biological interest. Note that replicate experimental units for each treatment are required, rather than sub-samples within a bulk of similarly treated material. A treatment must be repeated on several experimental units to establish its repeatability. It cannot be shown that an effect of a treatment is repeatable by sub-sampling from within a single application of that treatment. Papers which seek to demonstrate differences where there is no valid estimate of error may be rejected on the grounds that such claims must be regarded as unsubstantiated. In general, the more variable the material and the smaller the threshold at which effects are of interest, the more replication must be made. Arrangements of experimental units Experimental units should be grouped (blocked) according to known or suspected sources of variation. Examples are the grouping of animals to take account of variation in weight at the beginning of a growth experiment, or the selection of blocks of land of similar fertility when laying out a field experiment. After grouping, experimental units must be randomly assigned to treatments. This is discussed in most standard texts on design. Where data are collected by sampling from some natural population then a suitable strategy should be used so that comparisons of interest have valid estimates of error. Data analysis Types of data At least four distinct types of data could be collected in the course of scientific experiment or observation according to what is being investigated and the scale on which it is measured: (i) Quantities, such as weight or length, measured on an interval scale. Such data are said to have a continuous distribution. (ii) Variables measured on a semi-quantitative scale where observations can be ordered but the magnitude of the difference between successive points on the scale need not be constant. Examples are a score of an animal's body condition on (say) a ten point scale or a sensory judgement on (say) a four point scale from worst through to best. Under some circumstances (discussed later) the distributions of such variables might be approximately continuous but usually would not be so. Variables of this type are called ordinal categorical. An extreme example of this type is when data are ranks of observations. (iii) Qualitative observations classified into two or more categories. Examples are classification of animals by sex or a food tasted as (say) bland, bitter, savoury or sweet. These data would follow some discrete distribution such as a binomial (for two categories) or a multinomial (for more than two categories). Such variables are called nominal categorical. (iv) Counts of events occurring randomly in time such as radioactive disintegrations or cell divisions. Often such counts will follow a Poisson distribution. It is important to make these distinctions and to recognise in any given instance what type of data are under consideration so that an appropriate method of statistical analysis can be chosen. Analysis of variance Data measured on an interval scale (type (i) above) arising from designed experiments should be subject to an analysis of variance according to the design used and the questions of interest to the experimenter. Such analyses provide estimates of error from which standard errors for tables of means can be calculated. The separate calculation of individual standard deviations for each cell of a table is rarely justified except where there is evidence of heterogeneity of variance between groups or when means reported in the table are based on widely differing numbers of samples. Each variable in the table should, of course, have its own standard error. Structured treatments Where experimental treatments have a factorial structure, the breakdown of sums of squares in the analysis of variance should follow this structure. Where factors have quantitative levels, it may be helpful to use polynomial sub-models to explore any trends which might be present in the data (see references). Where treatments at first appear unstructured, it is nevertheless often possible to test comparisons, such as 'untreated control against average of all novel treatments' or 'treatments of a certain type against those of some other type', by means of orthogonal contrasts for which sums of squares can be calculated in the analysis of variance table. Use of analysis of variance in this way provides a much better basis for inference than does an arbitrary selection of t-tests between cells of a table of means. Comparisons between means Multiple comparison procedures (MCP), such as Duncan's multiple range test (DMRT), are valid (and then not necessarily appropriate) only where it can be assumed that the treatment means are uncorrelated and have equal variance. These procedures are not appropriate when the treatments have quantitative levels or factorial structure. Because MCP ignore structure, they cannot provide satisfactory interpretation of structured treatments. Even where MCP might be valid they are apt to give rise to paradoxes such as 'A is not significantly different from B', 'B is not significantly different from C', but 'A is significantly different from C'! Furthermore, the often overlapping groups of means rarely provide a natural framework for the interpretation of data and it is clear from the pages of the journal that results of tests (such as DMRT) although reported are seldom discussed. Authors are referred to the extensive literature on this topic for a fuller discussion of the problems with MCP and suitable alternative methods (see references cited). It is hoped that authors will find these alternative methods more useful. They may well be called upon to justify their use of MCP if alternative methods could be used to better effect. Least Significant Difference (LSD) is often used to make comparisons among a set of means. If used with care this can be a helpful informal technique in the exploratory phase of data analysis. Among many dangers is the temptation to make all possible pairwise comparisons. Other statistical considerations (see references) make this a statistic of little usefulness in supporting inferences from data. Authors are encouraged to use the more powerful and statistically valid methods mentioned in this paper. Observations repeated on the same material Where observations are repeated on the same experimental units at successive times, special methods of analysis may be required. This is due to the likely high correlation between successive observations made on the same experimental unit. For example, in an animal growth study the previous growth of an animal will very likely affect its future growth leading to correlated body weights at successive times of observation. See, for example, Rowell and Walters (1976) and Kenward (1987). Regression analysis If authors wish to establish the nature of a relationship between two or more quantitative variables then some sort of regression analysis is required. Simple analysis of variance with linear and quadratic orthogonal polynomials would suffice just to demonstrate trends in data. Regression analysis should never be done blind; graphical exploration of data is essential. When designing a study, care should be taken that the ranges of explanatory variables extend over useful intervals so that results will be as widely applicable as possible. The potential undue influence of isolated points is a danger. Examination of residuals from the fitted model will help in assessing how good the fit is. R2 (adjusted or otherwise) is of limited use in measuring goodness of fit because it measures the goodness of fit only where the relationship is linear and also it is sensitive to the design of the study. Standard deviations about the regression line will give a clearer picture as to whether the fitted model is adequate. See Atkinson (1987) and Cook and Weisberg (1985) for a discussion of regression diagnostics. Any non-linear models, such as exponential curves or logistic growth curves, should be fitted by least squares or the method of maximum likelihood using proven computer software (see below). Unlike simple linear regressions, these models require solution by iterative numerical methods. Analysis of multivariate data Where many distinct measurements on variables of a similar type (for example, absorbance at a number of wavelengths) have been made on each of the experimental units, authors are encouraged to consider use of a suitable multivariate analysis technique. These methods can sometimes give a much more concise summary of results than a set of univariate analyses in such circumstances. (See list of useful references.) Analysis of categorical data and counts of discrete events Ordinal categorical data Data which are scores on some (possibly arbitrary) ordinal scale may require special methods of analysis (see references). However, if the data can be considered as from approximately a grouped continuous distribution then analysis of variance may be used. A rough rule as to when this assumption could be justified is where the scale has about ten or so points and where the increments between points on the scale are similar in size. Nominal categorical data Observations which are classified into two or more qualitative categories give rise to data in the form of counts. These data could be analysed by non-parametric methods or by analysis of 2 on contingency tables. More sophisticated methods such as Generalised Linear Models (GLM) are described below. Statistical methods for data expressed as counts Data which are counts or counts expressed as percentages will usually require transformation to stabilise variance. Transformation of percentages is helpful where the range of data covers the extremes near 0% or 100%; it might not be necessary if the range is small and between 30% and 70%. A transformation which may be suitable is the angular transformation, sin-1 x/100 where x is a percentage and sin -1 means 'the angle between 0¡ã and 90¡ã whose sine is'. Also a suitable transformation may enable data to be modelled in terms of additive treatment effects and thus allow a simple explanation. For example, where the effect of a treatment was to cause a proportional increase in, say, counts of bacteria, a log transform would enable such effects to be modelled on an additive scale. The logit transformation (or 'log odds'), loge[p/(1- p)] where p is a proportion, may allow additive treatment effects to be fitted for data which are number responding out of a fixed total. For some data, where the twin objectives of stable variance and an additive scale for effects cannot be met by a single transformation, a Generalised Linear Model (GLM) could be useful. These GLM are like the familiar linear models of regression and analysis of variance but are more general in allowing error distributions other than normal, for example, Poisson or binomial, and data transformations to linearity (additivity) such as log, logit, complementary log-log, power, and several others. Generalised Linear Models are usually straightforward to interpret and are incorporated in most general statistical packages (see below). Statistical software There are many excellent statistical packages which perform most of the standard analyses. The following list is not comprehensive: BMDP, Genstat, Minitab, SAS, Splus, SPSS, Stata and Statistica. PRESENTATION Description of statistical methods Remember, unless your paper deals with new methodology or applications of statistics, that the statistics are supposed to be an aid to the interpretation of data, not an end in themselves. Include a clear, concise description of what you have done and only those details which you wish to refer to in the discussion of the results. It is rarely necessary to describe standard methods in full where reference can be made to a standard text. Equations may be helpful to make clear an adaptation of a standard method but all symbols used must be defined in the text. Begin with a description of the design of the experiment or investigation. State the objectives of the analysis and justify any underlying assumptions in the context of your application. Describe how the chosen method has been applied to these data. If a statistical package has been used to process data on a computer, give a reference to the manual. Making inferences from data Statistical analyses should be used to estimate effects and to justify inferences against a background of biological variation. They should be discussed and not ignored. Lack of discussion strongly suggests that the statistics has been included solely as a concession to try to ensure publication. The achievement of 'significant' differences from a statistical test should not be regarded as an end in itself! Indeed, sameness can be just as important as difference. Statistical significance is judged according to arbitrary probability levels and its attainment may well depend on the level of replication in an experiment. A statistically significant effect is not necessarily a biologically interesting one, nor vice versa. The estimated size of effects of treatments and their standard errors are often of more interest than is statistical significance. Tables Deciding on precision Do not be tempted to exaggerate the precision of results by giving too many decimal places. A useful rule of thumb is to round to approximately one tenth of the standard error and give standard errors with one more decimal place than this. Choosing how to give errors of estimation The use of '¡À' to indicate some sort of error is unsatisfactory, because usage varies in different countries. The type of error should be clearly stated; for example, standard error of mean (SEM), standard error of difference (SED), standard deviation (SD), and so on. The use of '¡À' will only be allowed where it is clearly stated, on first use, what is meant. As far as possible standard errors in tables should be based on pooled estimates of error. Separate standard errors for cells in a table are worthwhile only where there is evidence for heterogeneity of variance or where replication differs widely between classes. Presenting data transformed for analysis Where data have been transformed for analysis, results may be easier to comprehend if back-transformed. In this case confidence intervals, calculated on the transformed scale and also backtransformed, should be given for each entry in a table (although in general these are only approximate). There is little sense in back-transforming a standard error when the purpose of the transformation has been to stabilise variance. Use of Generalised Linear Models can provide summaries in terms of natural quantities, such as proportions, and so avoids problems of back-transformation. Avoiding superfluous statistical tests Tables should not be annotated with results of statistical tests unless these are referred to in the text of the paper. Routine annotation with superscripts according to Duncan's Multiple Range Test will not be acceptable (see above). Authors should themselves discuss all sensible inferences which can be drawn from their results; there is no need for any statistics other than those required to support these. Graphs and regression analysis Graphs are often helpful where regression results are presented. Error bars for standard errors of means or standard errors of prediction for fitted relationships are preferred. Do not use Least Significant Differences (LSD) for data where there is a quantitative structure to treatment levels as these should never be analysed by successive significance tests (see above for alternatives). The use of R2 (=sum of squares due to regression on x-variates ¡Â total corrected sum of squares for y-variate) to measure goodness of fit is discouraged and is unacceptable in the case of multiple regression as R2 takes no account of the number of terms in the model. Adjusted R2 (2=1-[residual mean square after fitting all x-variates ¡Â total mean square with (n-1) degrees of freedom]) or percentage variance accounted for (=100¡Á2) should be used instead as this does allow for the number of parameters estimated. Prior to the calculation of any summary statistic for goodness of fit, however, the residuals should be plotted and examined to detect any patterns which invalidate the model assumptions. ACKNOWLEDGEMENTS I am grateful to John Gower (AFRC), David Brown (AFRC), Clive Moncrieff (AFRC), Peter Rothery (NERC) and other colleagues for many helpful comments on earlier versions of these notes. USEFUL REFERENCES Good Statistical Textbooks Cochran WG, Cox GM, Experimental Designs. Wiley, New York (1992). Cox DR, Planning of Experiments. Wiley, New York (1992). Cox DR, Snell EJ, Applied Statistics: Principles and Examples. Chapman & Hall, London (1981). Draper NR, Smith H, Applied Regression Analysis (3rd edn). Wiley, New York (1988). Mead R, Curnow RN, Hasted A, Statistical Methods in Agriculture and Experimental Biology (2nd edn). Chapman & Hall, London (1993). Snedecor GW, Cochran WG, Statistical Methods (7th edn). Iowa State University Press, Ames (1980). Sokal RR, Rohlf FJ, Biometry (3rd edn). WH Freeman, San Francisco (1995). Steel RGD, Torrie JH, Dickey D, Principles and Procedures of Statistics. McGraw (1996). Weisberg S, Applied Linear Regression (2nd edn). Wiley, New York (1985). Some General Papers Chatfield C, The initial examination of data. J Royal Statist Soc A 148:214-253 (1985). Jeffers JNR, Statistical Checklist 1: Design of Experiments. Institute of Terrestial Ecology, Huntingdon, UK (1982). Jeffers JNR, Statistical Checklist 2: Sampling. Institute of Terrestial Ecology, Huntingdon, UK (1982). Preece DA, Good statistical practice. The Statistician 36:397-408 (1987). Polynomials and Contrasts Mead R, Curnow RN, Hasted A, Statistical Methods in Agriculture and Experimental Biology (2nd edn). Chapman & Hall, London, pp 94-97 (1993). Snedecor GW, Cochran WG, Statistical Methods (7th edn). Iowa State University Press, Ames, pp 224-228, 257-258, 307-315, 404-407 (1980). Sokal RR, Rohlf FJ, Biometry (3rd edn). WH Freeman, San Francisco, pp 229-259, 521-530, 665-677 (1995). Repeated Measures Kenward MG, A method for comparing profiles of repeated measurements. Applied Statistics 36:296-308 (1987). Rowell JG, Walters DE, Analysing data with repeated observations on each experimental unit. J Agric Sci 87:423-432 (1976). Regression Diagnostics Atkinson AC, Plots, Transformations and Regression. Clarendon Press, Oxford (1987). Cook RD, Weisberg S, Residuals and Influence in Regression (2nd edn). Chapman & Hall, London and New York (1985). Draper NR, Smith H, Applied Regression Analysis (3rd edn). Wiley, New York, Ch 3 (1998). Weisberg S, Applied Linear Regression (2nd edn). Wiley, New York, Chs 4, 5, 6 (1985). Multiple Comparisons Letter to Editors and short articles that concentrate on illustrating inappropriate published uses of multiple comparison tests and on explaining the alternatives Baker RJ, Multiple comparison tests. Canadian J Plant Sci 60:325-327 (1980). Bryan-Jones J, Finney DJ, On an error in instructions to authors. Hort Science 18:279-282 (1983). Little TM, Interpretation and presentation of results. Hort Science 16:19-22 (1981). Morse PM, Thompson BK, Presentation of experimental results. Canadian J Plant Sci 61:799-802 (1981). Pearce SC, Data analysis in agricultural experimentation. III Multiple comparisons. Expl Agric 29:1-8 (1993). Perry JN, Multiple comparison procedures - a dissenting view. J Econ Ent 79:1149-1155 (1986). Papers that either go into more detail about the principles and theory behind multiple comparison tests or are concerned with discussing the wider issues of experimental design and analysis, but include comments on multiple comparison tests Chew V, Comparing treatment means: a compendium. Hort Science 11:348-357 (1976). Chew V, Testing differences among means: correct interpretation and some alternatives. Hort Science 15:467-470 (1980). Cox DF, Design and analysis in nutritional and physiological experimentation. J Dairy Science 63:313-321 (1980). Multivariate analysis Barnett V (Ed), Interpreting Multivariate Data. Wiley, New York (1981). Everitt BS, Graphical Techniques for Multivariate Data. North-Holland, Amsterdam (1978). Krzanowski WJ, Principles of Multivariate Analysis: a User's Perspective. Clarendon (1990). Ordinal Categorical Data Agresti A, A survey of strategies for modelling cross-classifications having ordinal variables. J Amer Statist Assoc 78:184-198 (1983). Agresti A, Introduction to Categorical Data Analysis. Wiley, Chichester (1996). Fingleton B, Models of Category Counts. Cambridge University Press, Cambridge (1984). McCullagh P, Regression models for ordinal data. J Royal Statist Soc B 42:109-142 (1980). Generalised Linear Models Cox DR, Snell EJ, The Analysis of Binary Data. Chapman & Hall (1988). McCullagh P, Nelder JA, Generalised Linear Models (2nd edn). Chapman & Hall, London (1989). Adjusted R2 (2) Draper NR, Smith H, Applied Regression Analysis (2nd edn). Wiley, New York, pp 91-92 (1981). Healy MJR, The use of R2 as a measure of goodness of fit. J Royal Statist Soc A 147:608-609 (1984). Snedecor GW, Cochran WG, Statistical Methods (7th edn). Iowa State University Press, Ames, pp 358-360 (1980). Weisberg S, Applied Linear Regression (2nd edn). Wiley, New York, p 188 (1985). Editorial Board   EXECUTIVE EDITORIAL BOARD D S REID Editor-in-Chief (Food Science and Technology) University of California, Davis, USA e-mail: dsreid@ucdavis.edu R J DOWDELL (Agriculture: Production) Welwyn, Herts, UK e-mail:rodney.dowdell@soci.org M W GRIFFITHS (Food Safety) Department of Food Science University of Guelph, Ontario, Canada e-mail:mgriffit@uoguelph.ca P KROON (Biochemistry) Institute of Food Research, Norwich, UK e-mail:paul.kroon@bbsrc.ac.uk R B LESLIE (Special Projects and Developing Areas) Sharnbrook, UK jla@virgin.net P J LILLFORD (Biotechnology and Molecular Biology) University of York, UK pl8@york.ac.uk I M MORRISON (Agriculture:Utilization) Ayr, UK ianmmorrison@hotmail.com P J MOUGHAN (Food: Health and Nutrition) Massey University, New Zealand p.j.moughan@massey.ac.nz P RICHMOND (Materials and Processing) Department of Physics Trinity College Dublin, Ireland peter_richmond@totalise.co.uk M A SHEPHERD (Agriculture: Environment) ADAS Gleadthorpe Research Centre Nottingham, UK mark.shepherd@adas.co.uk EDITORIAL BOARD M J Adams Royal Melbourne Institute of Technology Australia e-mail: mike.adams@rmit.edu.au S Arai Department of Nutritional Science Tokyo University of Agriculture Japan J M Besle INRA de Theix, St Genes-Champanelle France e-mail: besle@clermont.inra.fr L F Bisson University of California, Davis, USA e-mail: lfbisson@ucdavis.edu N Blakebrough Cheshire, UK e-mail: norman.blakebrough@ntlworld.com R Booij Plant Research International Wageningen The Netherlands e-mail: r.booij@plant.wag-ur.nl M P Cano Instituto del Frio - CSIC Ciudad Universitaria Madrid, Spain e-mail: pcano@if.csic.es M N Clifford University of Surrey Guildford, UK e-mail: m.clifford@surrey.ac.uk P Colonna Departement de Transformations ¨¢¨¢des Produits Vegetaux Institut National de la Recherche Agronomique, Nantes, France e-mail: colonna@nantes.inra.fr J S Conway Royal Agricultural College, Cirencester UK e-mail:john.conway@royagcol.ac.uk P Dalev University of Sofia Bulgaria e-mail: pdalev@biofac.uni-sofia.bg H Dove CSIRO, Canberra, Australia e-mail: h.dove@pi.csiro.au M Ellis University of Illinois, Urbana, USA e-mail: m-ellis7@ux6.cso.uiuc.edu L Francis AWRI, Glen Osmond, Australia e-mail: leigh.francis@swri.com.au C Geissler King's College London, UK e-mail: catherine.geissler@kcl.ac.uk G M Hall Loughborough University of Technology UK e-mail: g.m.hall@lboro.ac.uk J H Hotchkiss Cornell University, Ithaca, USA e-mail: jhh3@cornell.edu N K Howell University of Surrey, Guildford, UK e-mail: n.howell@surrey.ac.uk F Jessen Ministry of Food, Agriculture & Fisheries, Technical University of Denmark, DK-2800 Kgs. e-mail: flj@dfu.min.dk I T Johnson Institute of Food Research Norwich, UK A A Kader University of California Davis, USA e-mail: aakader@ucdavis.edu R L¨¢sztity Technical University Budapest Hungary e-mail: lasztity.bet@chem.bme.hu G R List National Center for Agricultural Utilization Research, Illinois, USA e-mail: listgr@ncaur.usda G R List National Center for Agricultural Utilization Research, Illinois, USA e-mail: listgr@ncaur.usda I Lumley Laboratory of the Government Chemist London, UK e-mail: idl@lgc.co.uk J A Marlett University of Wisconsin-Madison USA e-mail: jmarlett@plantpath.wisc.edu A Mascarenhas-Ferreira Universidade de Tras-os-Montes e Alto Douro Vila Real, Portugal e-mail: amf@utad.pt J Ralph USDA-ARS Madison Wisconsin USA e-mail: jralph@facstaff.wisc.edu D Simatos Universit¨¦ de Bourgogne Dijon, France e-mail: dsimatos@mail.u-bourgogne.fr R Tester Caledonian University Glasgow, UK e-mail: r.f.tester@gcal.ac.uk E Tornberg Kemicentrum Lund, Sweden e-mail: eva.tornberg@livstek.lth Y van der Honing ID-DLO Leylstad, The Netherlands e-mail: y.vdhoning@hccnet.nl A L Waterhouse University of California Davis, USA e-mail: alwaterhouse@ucdavis.edu R Wilson Institute of Food Research Norwich, UK e-mail: reg.wilson@bbsrc.ac.uk W J Wolf USDA, ARS, Peoria, Illinois, USA e-mail: wolfwj@ncaur.usda.gov K Yasumoto Kyoto University, Japan