Appendix A - WRITING SCIENTIFIC PAPERS IN BIOLOGY

In most courses in the Biology Department, you will be asked to write papers that describe projects you have performed in the laboratory. Like papers in non-science courses, we expect them to be well-written, creative and thoughtful. Unlike papers in other courses, we expect them to conform to a format common to biological journals, which is described below. It is critical that you understand the details of this format, and particularly that you do NOT confuse these papers with "lab reports" required by other science courses at Grinnell.

The following describes how your paper should be organized. However, this doesn’t mean that you should write the paper in this order! We suggest writing the paper in this order: (1) Materials and Methods (2) Results, (3) Discussion, (4) Introduction and (5) Title.

In each section below are listed some "fatal flaws." If we find these in your paper, we’ll return it for a rewrite, with significant penalty.

The title tells what the paper is about. Under it should be your name and "professional address," which in this case is the specific lecture and laboratory sections in which you are enrolled. For this course, it is sufficient to put this at the top of the first page, although other professors may ask you to have a separate title page (We don’t in order to save paper).

A title should be informative, specific and concise. Since you are not writing a murder mystery, it is all right to tell the "ending" in the title. It is often this information that helps a reader decide if the paper is something he/she wants to read. Here are few examples that illustrate this:

• Really Bad Title: "Lab 2 - Plant phenotypes " This tells the reader nothing except that the second lab of the semester involved looking at plant phenotypes! Although it may distinguish this project from others in this course, it would not help someone outside of class understand what the paper is about. This is a fatal flaw.
• Better, but not very good: "Growth and phenotypic variation in Solidago gigantea" This gives specific information about the organism and type of measurement, but still does not inform the reader about the problem being investigated. Why do you want to measure variation in growth of goldenrods?
• Good: "Solidago gigantea growth shows lower heritability than morphological characters " The title indicates that the objective was to compare levels of heritability in different characters of a goldenrod, and even what the result was!

The introduction should briefly explain why the research was done, relating this research to other relevant work and giving the reasons for choosing the hypothesis to be tested. A good way to organize the introduction is to begin with the general and proceed to the specific. Assume that the reader is at least moderately familiar with the general subject of the paper. However, unless you are studying a model organism (e.g., Drosophila, Arabadopsis, E. coli etc.), it is important to describe enough aspects of its natural history that the reader can appreciate why it was chosen for the study. A good introduction will mention the major issues that will be considered in the Discussion section, and that’s why we suggest writing it after finishing the other sections.

Fatal flaw: Don’t start the paper with the phrase: "The goal of this experiment was . . . " This is not a general statement about the subject.

This section should carefully explain how the research was done. Organize the sections logically, using subheadings if there are more than several paragraphs. Include all materials used, the exact conditions employed, and how you gathered the data. You may cite the lab manual (or another source) for a common technique. Here’s an example:

• Bad: "We measured growth rates and wrote down the data." How did you measure growth rates? You don’t need to tell the reader that you recorded the data or entered it into the computer.
• Good: "We estimated yeast density according to directions given in the Sources of Genetic Variation lab handout (Brown and Eckhart, 1997)."

If you develop your own technique, you should explain it in sufficient detail that another person could replicate your work. Tell exactly what materials you used (composition of solutions, media, etc.) and describe the procedures precisely. Commonly used statistical tests generally need no explanation or citation.

Fatal flaw: Don’t present your materials and methods as a list. Write in complete sentences and organized paragraphs.

The results section should summarize, but not interpret, the results obtained. One good way to approach the writing of the Results section is to develop a set of questions about the data you gathered. Don’t use any questions that begin with "Why" -- these necessarily involve interpretation, so they should be addressed in the Discussion section. Write your Results section by answering each of these questions in a logical order. Refer to Figures and Tables as you describe the results.

Figures and tables frequently will help the reader to understand more easily than a written description. They should not duplicate text, and text should only tell the reader the major points to be noted on the graphs or tables (that is, should tell the reader what you think is particularly important about the data presented). Obviously the same data should not be presented in two different forms, so decide which form helps you tell your reader what you want him/her to know. Graphs of any kind, as well as other pictorial materials, are referred to as "Figures" in the text, and are numbered. Tables are called "Tables" in the text, and are numbered separately from figures. Each figure or table should be on a separate page. They may either follow the first reference to them in the text, or all may be collected at the end of the paper. All figures and tables must be "called out" or referred to before they are explained in the text.

• Bad example: "The results are shown in Graph 1." This is not a summary of the results, and the graphs should be referred to as "Figures." This is a fatal flaw.
• Another bad example: "Growth rates under low fertilizer had an average of 3.2 g with a 95% confidence interval of ± 1.4 g, and growth rates under high fertilizer treatment had an average of 6.5 g with a 95% confidence interval of ± 1.0 g (see Figure 1). Since the intervals do not overlap, they are significantly different at the 0.05 level." The text here repeats specific information that is also shown in the figure -- since the figure conveys this information more efficiently it’s not necessary to repeat it in the text. It is also not necessary to explain why non-overlapping confidence intervals indicate a significant difference, since this is a commonly used statistical test.
• Good example: "We found that higher levels of fertilizer resulted in significantly larger growth rates, as determined by the 95% confidence intervals (Figure 1)."

• Use a graph to illustrate a relationship or pattern in your data.
• Be sure that the type of graph you choose is appropriate for the type of data you wish to display. What assumptions are you making by using a particular type of graph? For example, a line graph with the points connected indicates that the variables are continuous over the range displayed.
• By convention, the independent variable (the one you manipulated) is on the X-axis and the dependent variable (the result of the manipulation) is on the Y-axis.
• The axes of a graph must have clear, concise labels. If there is more than one line or bar on the graph, each must be clearly identified.
• All figures must have clear and specific legends. A legend is usually written as an incomplete sentence with only the first word capitalized and should begin with "Figure 1. The . . ." If required for clarity, you may include several more sentences. It should be placed above or below the figure. [Note for users of EXCEL: In EXCEL’s very nice "Chartwizard", a "legend" refers to the part of the graph that explains what various lines/colors refer to. Use the "Title" box in the Chartwizard to type in a figure legend.]

Imagine that you have done an experiment that compares growth rates of Brassica under two different fertilizer treatments. The following is an example of a BAD GRAPH:

This graph makes poor use of space (note the big empty area). More seriously, the reader can’t tell from the figure whether the bars are mean values, and whether the difference between them is significant -- there needs to be some indication of the variation within each treatment group, i.e., error bars. Note as well that the Y axis has no units and the figure legend is not explanatory.

Here’s a much better version of this graph:

• Tables are used to present matrices of data. If it is important to show a pattern or trend, use a graph instead of a table.
• Do not present raw data and expect your reader to do the arithmetic before s/he can understand the contents of a table. Printouts of EXCEL worksheets with raw data are NOT acceptable tables! This is a fatal flaw.
• Try to avoid large tables - no one will read through them. Perhaps the information can be presented better in several smaller tables.
• All tables must have legends that explain their contents sufficiently that they can stand alone (much like figures). It is sometimes appropriate to have footnotes for a table.

The discussion should interpret and explain the meaning of your results and usually proceeds from the specific to the general. Begin with a summary of your results in a sentence or two. Remind the reader of important trends, etc. Then relate your results to your own initial hypothesis, arguing for a particular interpretation of your data. Don’t forget that "negative" results can be important too, since they may suggest that your hypothesis was incorrect (After all, it is just possible that you did everything perfectly!). Relate your results as well to other papers or published hypotheses if appropriate.

The discussion should end with a summary - the "take home lesson" that you want your reader to remember about your work. It may also raise further questions for study. However, if you end the discussion with the phrase, "but of course more work needs to be done," we will lower your grade! Please let me know what kind of work would be most informative and why.

Acknowledgments are used to thank any persons who contributed any significant help during the study. Such contributions include, but are by no means limited to, help in experimental design, collection of data, preparation of graphs, drawings or the manuscript, critiquing a draft of the manuscript or in financial or physical support of the work. Your partners in a group project must certainly be acknowledged!

The references list only the papers or other publications that were directly referred to in your paper; they are not a bibliography. References should be cited in one of two ways in the text: (1) Mention the authors' names as part of your sentence followed by the year of publication in parentheses; (2) Place authors' names and the year of publication in parentheses following ideas or results from the study. See examples below:

(1) "Brown and Wilson (1992) performed a set of experiments demonstrating variability in host specialization . . . "

(2) A phylogenetic tree of the yucca moth family indicates the important role of host-plant association in the evolution of this group (Brown et al. 1994)."

The following are examples of how articles should appear in the References section). Articles should be listed in this section alphabetically according to the first author.

Journal article: Brown, J.M. and D.S. Wilson. 1992. Local specialization on sympatric hosts: phoretic mites on carrion beetles. Ecology 73:463-478.

Book: Brown, J.M. 1999. How I became rich in academia -- a fantasy. Harper and Row, New York, NY.

• Biological editors recommend using the active, rather than the passive, voice in scientific reports. Thus, instead of writing, "The plants were measured . . .", you should write, "We measured the plants . . ." Note that this convention may not be true in other disciplines (e.g., chemistry).
• Use the past tense when describing your experiment and its results. Use the present tense when discussing general properties of organisms.
• No hand-written papers accepted.
• All text must be double-spaced with 1 inch margins on all sides.
• Pages must be arranged in the correct sequence, numbered and stapled together.
• Bear in mind Johnson’s comment on Milton’s Paradise Lost: "If its length be not virtue, it hath no other." Every word costs money to print, so keep it as short as possible consistent with a good job.
• And finally, never apologize for lack of data; assume that you have done the best job possible!

All papers will be reviewed as if we were doing it for a regular journal. The evaluations will be given as follows:

"Accepted" - a paper that could be published without revisions (an A paper, 45-50 pts.)

"Accepted with minor revisions" - a paper that could be published with attention to a few comments (a B paper, 40-44 pts.) . Criticisms may include

• a minor error in way data are presented,

• a minor error in data analysis,

• a lack of clarity, especially in the discussion,

• awkward or improper sentence construction.

"Requires major revisions" - a paper that is not publishable without extensive work (a C or D paper, 30-39 pts.). Criticisms may be

• inadequate data presentation and/or analysis, which leads to

• faulty conclusions,

• lack of creative thought in discussion,

• poor writing style or lack of proofreading and spell checking.

"Rejected" -- must be rewritten to receive a grade. Due to

• a fatal flaw
• a lack of understanding of the premise of the experiment,
• or a failure to complete a section of the paper.

We accept late papers with a penalty of 5 pts/24 hours, or portion thereof. No penalties will be assessed with evidence of illness from the Health Center.

(Adapted from Writing Papers in the Biological Sciences (1988) by Victoria E. McMillan, St. Martin’s Press, N.Y. Grading scheme adapted from Bill Wirtz’s handout for Biology 112 , Pomona College)