COURSE MANUAL
BIOL 3630: FIELD BIOLOGY
FALL 2011
Department of Biological Sciences
University of Lethbridge
Instructor: Dr. Cam Goater
WE - 1048
cam.goater@uleth.ca
1
1
TABLE OF CONTENTS
OVERVIEW………………………………………………………………………………………3
FIELD BIOLOGY 2011 ................................................................................................................. 3
SCHEDULE OF ACTIVITIES…………………………………………………………………...4
A NOTE ON HAZARDS ............................................................................................................... 5
GRADING SCHEME ..................................................................................................................... 5
SPECIFIC PROCEDURES ............................................................................................................ 6
Sampling Arthopods ................................................................................................................... 7
Point Frame Analysis of Plant Abundance……………………………...………...……………7
Plant Biomass………………………………………...………………………………………...8
Mark-recapture estimates of animal populations……………………………………………….8
Intraspecific competition in Lodgepole Pine………………………………………………….10
WRITING SCIENTIFIC REPORTS ............................................................................................ 11
ORAL PRESENTATIONS ........................................................................................................... 16
THE USE OF SIMPLE STATISTICS IN FIELD BIOLOGY ..................................................... 18
APPENDIX ................................................................................................................................... 27
Table of Random Numbers ....................................................................................................... 28
Natural history of the Goldenrod ball gall – the key players …………………………………29
Distribution of parent and hybrid cottonwoods in Southern Alberta………………………….30
Leaf morphology of parent and hybrid cottonwoods ………………………………………... 31
2
OVERVIEW
The first aim of this course is to provide you with the opportunity to obtain hands-on experience
outside the classroom. First and foremost, such an opportunity forces you to confront and
recognize the real-world complexity of biological systems. Embedded in this principle is the
idea that ‘you learn a lot by doing’. A second aim is for you to develop skills in research
methodology within a field setting. Performing experiments under natural or semi-natural
outdoor conditions presents several challenges and addressing these with properly designed
experiments is an important skill to learn. So we work on your ability to go into the field, to pose
workable problems, design appropriate experiments/studies, and then collect appropriate data.
Another aim is to develop your skills at applying simple statistical tests for the analysis of fieldcollected data. Lastly, we work on developing your skills at scientific communication, both in
written and oral format. The former parallels what you would see in a modern journal, the latter
is what you would observe at a national meeting of ecologists or field biologists.
FIELD BIOLOGY 2011
The format of the course is unconventional. The Field Trip in Cypress Hills Park (21-27 August)
is a mandatory component, comprising a series of dedicated experiments requiring input from all
students. Each experiment is data-intensive. Although we will complete as many of the formal
analyses as possible during the field trip, some will carry-over into the regular term. In addition
to the formal experiments/studies, I (or an invited guest) will also provide impromptu minilectures at various locations and times throughout the Field Trip. In September, the class will
meet on Tuesday/Thursday (10:50 –12:05) in WE-2084 for data analyses and interpretation.
There are no formal lectures following the Field trip.
3
TENTATIVE SCHEDULE OF ACTIVITIES
Location
Camp McCoy
21 Aug
7:00-9:00 pm
Topic
Study
type
Analyses
Introduction and welcome
22 Aug
8:30-12:00
12:00-1:00
1:00-4:00
7:00-8:00
9:00-
CHP plateau
Reesor Lake
CHP plateau
Camp McCoy
CHP plateau
Disturbance ecology
Prairie upland entomology
Dispersal/competition trade-offs
Seminar: Kevin Floate1
Forest/grassland entomology
FE
FS
FE
23 Aug
9:00-12:00
1:00-4:00
7:00-8:00
9:00-
CHP plateau
Various sites
Camp McCoy
Camp McCoy
Disturbance ecology
Mark-recapture
Seminar: Melissa Thomson2
Data collation
FE
FS
24 Aug
9:00-12:00
1:00-4:00
7:00-9:00
9:00-
CHP plateau
CHP plateau
Camp McCoy
Camp McCoy
Competition; Parasitism
Forest entomology (Les Weekes)
Research proposal presentations
Data collation
25 Aug
9:00-4:00
7:00-10:00
Various sites
Camp McCoy
Independent projects
Data collation
27 Aug
9:00-4:00
7:00-10:00
Various sites
Camp McCoy
Independent projects
Data entry and organization
Regr/corr; Anova
Regr/Anova
FS
Regr/corr; Anova
Model application
Regr/model test
FS
FE
FE; FS TBA, various
FS
TBA, various
TBA
28 Aug
9:00-12:00
Class hike
FS = Field Study; FE = Field Experiment
1. Kevin Floate (Agriculture and Agri-Foods Canada): Cottonwood/insect interactions in
southern Alberta
2. Melissa Thomson (M.Sc. candidate, U of L): Epidemiology of an invasive parasite in CHP.
3. Kim Dohm (Ph.D. candidate, U of L): Biology of sprague’s pipit on native and disturbed
grasslands.
4
A NOTE ON HAZARDS
There are hazards associated with every trip that occurs outside the walls of the University. The
approach that I take regarding safety is one that emphasizes common sense and recognition of
the fact that you are adults. Throughout the 6 days of field trips, we will be traveling to various
local sites. Hazards associated with being outside, such as weather extremes, exposure to insects
and other animals (e.g. rattlesnakes, scorpions, wasps and bees, bears) are significant features at
several of these sites. You must come prepared with appropriate foot wear and other clothing,
and with sun screen and sun hats. Bug spray is also advisable. In general, we do everything we
can to minimize risks associated with field work, but they are still present. Obviously, drinking
alcohol is not allowed during the portion of the field trips that can be considered part of the
course. I will discuss specific potential hazards when we arrive at the various sites.
COURSE SYLLABUS
Grading Scheme:
Exam
Written report I
Written report II (First draft)
Written report II (Final draft)
Oral presentation
25
25
15
25
10
Exam:
The material for this exam will come from your field notes and from
collections and images taken during the field trips. Each of the minilectures and themes covered during the field trips will be associated with
specific papers and review articles drawn from the primary literature.
Copies of these supplementary readings will be made available. Details
on the general format of the exam will be provided at the beginning of
term. There is no final exam.
Written report I (Class data):
Students will write a report following standard format for scientific
communication. This report will involve the analyses and interpretation of
selected data sets collected by the class during the field trips in CHP. See
handout for details regarding format.
Written reports II (Research project):
The independent research project is an important component, worth 50%
of your overall grade. Students will write an initial draft. You will receive
comments and suggestions for improvement on this draft, which you will
then use to complete a final draft. The reports will follow standard format
for scientific communication.
Oral Presentation (Research project):
Students will provide 15-minute oral summaries of their Research
projects. The first aim is to provide experience in the development and
5
presentation of oral presentations. The second is to allow for your
classmates to provide feedback prior to the due-date for completion of the
final draft.
Course grades:
Final grades will be based upon performance in the various portions of the course in the
proportions given above. The following definitions and percent values are those that will be
used in determining grades, including final letter grades:
A+
A
A-
excellent
excellent
excellent
91% to 100%
85% to 90%
80% to 84%
C+
C
C-
satisfactory
satisfactory
satisfactory
67% to 69%
63% to 66%
60% to 62%
B+
B
B-
good
good
good
77% to 79%
73% to 76%
70% to 72%
D+
D
F
poor
minimal pass
failure
55% to 59%
50% to 54%
<50%
Policy on late reports and missed exams:
I must be notified in advance if an assignment will be handed-in late. Prior approval requires a
written notification giving detailed rational at least 3 days in advance of the deadline. Approval
will only be granted in exceptional circumstances. Late assignments will be penalized at 10% per
day, up to 3 days.
Schedule and important dates:
Date
Topic
21-27 Aug.
Sept. 8, 13, 15, 20
22 Sept.
29 Sept.
13 Oct.
22 Oct.
30 Oct.
Field Trip – Cypress Hills Provincial Park
Data collation and analyses
Exam (WE-2084)
Written report I due
Written report II (First Draft) due
Oral Presentations (Research project)
Written report II (Final Draft) due
SPECIFIC PROCEDURES
General
We will discuss in detail the specific methodologies associated with each of the class-based
studies. The following brief descriptions provide a background to some of the specific
methodologies that are repeated between studies, or are slightly non-conventional in their
approach.
6
A. Sampling Arthopods
Some of our studies involve comparing Arthropod and plant biodiversity between different
‘treatments’. The treatments are grazing, or fire, or mowing. Regardless of the treatment, the
methods we will use are approximately the same. The areas selected for sampling within a
treatment (termed “blocks”) will be approximately 50 m by 50 m. The corners of these areas or
blocks will be marked with flagging tape. Within each site, 10 plots will be selected using a
random number table. Each plot will be used to collect arthropod and plant biodiversity data.
The treatment block can be considered a grid with an origin at the south-west corner. Any point
within the grid can be identified by an east-west coordinate (or the distance moving east from the
origin [south-west corner] of the treatment block) and a north-south coordinate (or the distance
moving north from the origin [south-west corner] of the treatment block). When choosing a
random location for the placement of the quadrat, use a random numbers table to select an eastwest coordinate and a north-south coordinate. Then move to that position on the treatment block
and set up your plots.
Arthropods will be collected using pitfall traps to collect ground-dwelling types, and pantraps to
collect pollinators and fliers. At each plot, 2 pitfall traps will be dug into the ground, and 2
yellow pantraps placed in close proximity. Each plot should be arranged within a 1 m radius.
Sweep nets will also be used to sample the larger, flying insects. One pass with a sweep net
involves 10 ‘sweeps’ as you walk in one direction away from the centre of the plot, sweeping the
grass as you walk. Do 4 passes per plot, sweeping in N, S, E and W directions away from the
centre. All of the arthropods from the 4 sweeps can be placed into one, labeled container. After
24 hours, we will re-visit the plots and remove the arthropods from the pan and pitfall traps.
Samples from the 4 traps and the 4 sweeps will be pooled prior to analysis.
B. Point Frame Analysis of Plant Abundance (Percentage Ground Cover)
There are several methods that can be used for determining plant abundance and cover. In this
course you will be introduced to the point frame method of determining the percentage of ground
covered by plants in an area. Metal pins with sharp tips serve as the points. The pins are arranged
in frames that rigidly limit the pin to a vertical path perpendicular to the ground. The point frame
is used to sample vegetation by placing the frame a specific position on the ground and lowering
a pin until it first touches a plant. The plant species touched by the pin is recorded, and then the
pin is lowered further until it touches another leaf, and so on until the ground is reached. The
data collected allow the calculation of the percentage ground cover occupied by a particular plant
species at a site:
% cover 
Number of pi ns whi ch hi t speci es A
x 100
Tot al number of pi ns
Theoretically the point frame pins represent a sample of the infinitely large number of points
within a given area, and it is because of this fact that the smaller the point the more accurate will
be the % cover estimates. Similarly the larger the number of point samples, the better they will
represent the true vegetation abundance. The point frame method is a very objective method for
7
determining vegetation abundance. However, it is usually only applicable to very low stature
vegetation such as grasslands, or forest understory vegetation.
Since a single point frame has only 10 points it is necessary to make numerous placements of the
point frame in an area to get a representative determination of plant cover. The exact method for
replicate placements varies depending on the objective of the investigation. In our case, we will
use the Arthropod plots as random positions to determine % cover. Thus, at each plot (each
replicate), place the Frame 2 m E and W, respectively from the centre. Each true replicate will
thus consist of 20 points sampled at each plot. At each placement of the point frame you should
note the number of times a point “hits” a specific plant species and also note “hits” on bare
ground. Data should be recorded on the data sheets provided.
C. Plant Biomass
Measurement of the percentage of ground covered by plants gives a simple and quick indication
of plant abundance that is useful in analyses of plant community composition and diversity. Plant
percentage cover should also be related to the productivity of a given site. However, depending
on the types of plants present (trees, shrubs, grasses), similar values of plant percentage ground
cover can represent very different values of biomass production. Primary production is an
important ecosystem characteristic that also affects many other ecosystem properties. The
conversion of light energy to chemical energy during photosynthesis and subsequent production
of plant biomass provides the energy source for all other trophic levels in an ecosystem.
Therefore in many studies it is more useful to directly measure biomass and primary production
than percentage ground cover. In this exercise you will be introduced to some simple methods
for determining above-ground plant biomass values.
Measurements of above-ground biomass will be made in treatment and control sites. Biomass
data are gathered by clipping vegetation at ground-level in randomly placed quadrats. The size of
the quadrats used in grassland sites is usually 20 cm x 50 cm, or 50 cm x 50 cm. Optimum
quadrat size is a function of the spatial variation in biomass over the study area and the time
required for harvesting and processing samples. Your quadrat should be placed at a position
determined at random from the centre of the plots (e.g. 2 m north, 1 m west). Harvest the plant
biomass (live and dead vegetation) and place it in marked paper bags. Bags should be marked
with your name, the date, sample type (eg. above-ground biomass), the site (eg. grazed grassland,
ungrazed grassland, etc.). Upon return to the lab, the paper bags should be placed in a drying
oven at 60°C. After drying, the live and dead tissue need to be weighed on a balance. Weight
measurements should be done using the Mettler PJ400 balance. We will use the data collected to
compare the effects of grazing on plant biomass. Above-ground live and dead plant biomass data
should be expressed in units of g m-2.
D. Mark-recapture estimates of animal populations
Estimating animal population size is one of the most important elements of studies in Field
Biology. But direct counts of numbers per unit area or unit volume are often impractical. In the
case of mobile or secretive animals, it is difficult to obtain direct estimates of the density of
individuals, even in areas of small size. Many of these populations are the exact ones for which
8
resource managers require accurate counts in order to set limitations on catch. One way to
resolve this problem is to mark a segment of the population on one occasion, and sample the
ratio of marked to total animals on one or more later occasions. We will use one of the many
mark-recapture techniques (Petersen or Lincoln Index) to estimate the population sizes of
terrestrial snails in CHP. This exercise is relevant to the methodology component of the course,
to the long-term monitoring component, and to the applied component.
The Petersen Index is not just an index of relative population size, but a true estimate of the
population size of the area sampled. In this technique, a sample of snails is captured, marked,
and released back into the population, creating a certain ratio of marked to total animals. After
an interval sufficient to permit dispersal of the marked snails throughout the population, a second
sampling is carried out to estimate this ratio. From the size of the marked sample, M, the total in
the second sample, C, and the number of marked fish recaptured in the second sample, R, the
size of the population, N, is estimated by the equation,
N = MC
R
If the number of marked animals recaptured is expressed as a proportion, p, of the total in the
second sample, the equation becomes,
N
= M
P
This equation shows that the ratio of marked to total fish in the recapture sample, and not the
total number of recaptures, determines the population estimate.
In practice, there are a number of assumptions of this methodology that restrict its utility,
1. All individuals in the area must be equally catchable
2. The ratio of marked to total fish must not change between the time of release and recapture
(i.e. no recruitment into the population via births or immigration).
3. No loss of marks and no differential loss of marked vs unmarked fish by death or emigration
Analysis
Fill out the following table,
Number marked, M
Number recaptured, R
Total in second sample, C
Population estimate, N
p (=R/C)
q (1.0 – p)
Confidence limit for p
Confidence limit for N
Confidence limits are easily calculated. To do so, confidence limits for p, the proportion of
marked fish in the recapture sample. The upper, pu, and lower, pl, confidence limits on p are,
9
95% CI pu = p + 1.96 pq, pl = p – 1.97 pq
C
C
Where p = R/C and q = 1 – p. To obtain CI for N, M is divided by the values calculated for pu
and pl.
E. Intraspecific competition in Lodgepole Pine
Ecological interactions such as competition are often studied under controlled conditions. These
controlled environment studies have provided many important insights into the mechanisms of
competition, and how differences in competitive ability among species can depend on resource
availability. However, for long-lived species such as trees, short-term, controlled experiments
may not be that meaningful. In order to provide insight into processes influencing the ecological
interactions of trees, experiments would have to continue much longer than the lifetime of most
investigators. It is possible, however, to study the effects of some interactions by examining the
patterns they produce. For example, widely spaced trees should compete less than closely spaced
trees, so that widely spaced trees should be larger than closely spaced trees, assuming that all
other processes are held constant. Thus, the sum or total size of two nearest neighbour trees tend
to increase with the distance between neighbours.
Work in groups of two. Each pair needs a tape measure and a notebook. Make a table in your
notebook with six columns:
Pair# CBH of Tree Basal Area
Distance to Neighbour CBH of Neighbour
Basal Area
Each group will be assigned to a particular study block within the forest. Randomly choose a tree
within the assigned block. Measure the tree’s circumference at breast height (CBH, 1.35 m above
ground) in cm. Measure the distance to the tree’s nearest neighbour in cm. Measure the
neighbour’s circumference in cm. Enter these measurements into the table in your notebook.
Repeat the measurements for 30 pairs of trees within your block. If a randomly chosen tree is
dead, select the closest live neighbour. Never measure the same tree twice. If a previously
measured tree is selected, move to the nearest unmeasured tree.
For analysis, calculate the cross-sectional (basal) area (m2) of the tree stems from the
circumference measurements. Determine the sum of the areas for each pair of trees. Data will be
combined from all groups for analysis later. Then plot the tree basal area (m2) versus the nearest
neighbour distance (m) to analyze for the influence, and intensity of competition.
Circumference = 2πr
Area of a circle = πr2
10
WRITING SCIENTIFIC REPORTS
In a nutshell, scientific communication is one of the most important skills that you can learn at
University. Yet effective writing is also a difficult skill to master, and it takes years of practice.
One of the problems for Biology students in particular is that there is often ‘much writing to be
done, but little time to focus on doing it well’ (Pechenik, 2001). The aim of this appendix is to
provide you with some advice on the mechanics and pitfalls of scientific writing. My sources for
this advice come from Pechenik (2001; this superb book is available upon request), a handout
written by Gail Michener for her class in Animal Physiology, and from notes taken from a
graduate course I taught at U of Alberta on Scientific Communication.
For students in 3rd or 4th yr University, poor reports usually stem from one (or a combination) of
4 problems. The first is that many of you have some training in writing ‘essays’ or ‘opinion
papers’, and have difficulty in changing to the ‘science’ style. The aim of Scientific Writing is to
be concise and clear, not verbose and flowery. Second, many students simply have poor writing
skills. The simplest rules involving nouns, adverbs and clauses can go a long way to making
your paper clear and easy-to-follow. Third, unclear writing usually stems from unclear thinking.
Usually, if you are struggling with ‘Writers Block’, it means that you are not exactly clear in
what you want to write about. Thus, one excellent way to truly ‘learn’ Biology, is to write about
it. Lastly, it is my experience that many students do not take the task of writing seriously.
Your lab reports are to be written in a style typical of that required by scientific journals that
report original research. The purpose of scientific writing is to convey information and ideas
exactly, explicitly, and economically. If this is your first attempt at writing a formal report, you
should pay particular attention to the instructions provided here.
Title
In a nutshell, the title should summarize what lies in the Introduction and Results sections. The
aim is to grab the interest of potential readers, right from the start. Avoid non-informative titles
such as ‘Zooplankton of Tyrrell Lake’, or ‘Feeding preferences of fathead minnows’. Replace
with something like ‘Species composition of fall-collected zooplankton in a prairie lake’ and ‘Do
fathead minnows, Pimephales promelas, select particular prey?’ The title should be on a separate
page, with your name (do not include your student ID), email, affiliation and date at the bottom.
Abstract
The aim here is to summarize in a few (5-8) sentences the major points of the study. You should
write it last. As a first step, start with a sentence that covers the general problem/phenomenon
being considered (e.g. Disturbance events are well known to affect arthropod community
structure). Next, shift to a mention of the system being studied (e.g. The effects of disturbance
on prairie arthropods were studied ….), then a brief mention of the techniques used, then the
most important findings. Always conclude with a general statement that covers the principal
conclusions reached (e.g. These results show that spatial and temporal variability in parasitoid
infection of gall formers is due to variation in host density). The abstract is always written in the
passive voice. Also, be sure to make your abstract informative. For example, the sentence
‘Grazing affected arthropod biodiversity’ is probably true, but it is not very informative. Try
11
something like ‘Grazed sites contained 34% fewer species and 42% fewer individuals than
control sites’.
The abstract is important, but it is notoriously difficult to write. Often, it is the only part of your
paper that your intended audience will read. Don’t leave it to the 5 minutes before your paper is
due.
The abstract, together with the rest of your paper, should be double-spaced and written on one
side of the page. Number all pages.
Introduction
The aim is to orient the reader to the general nature of the problem under consideration. Because
the experiments are designed to answer questions, the Introduction should inform the reader
about what questions are being investigated and why those questions are of biological interest
and importance. The Introduction is often the most difficult component of a paper to write
properly. The problem is that to encourage a reader to keep going, you have to convince him/her
early-on, that you are an authority on the question of concern. This is not easy for young
researchers, because it requires extensive background reading and clear understanding of the
topic.
As one suggestion, aim to think in terms of paragraphs. For most papers that you will be writing
for me, you will never need more than 3-4 paragraphs to get your message across. Try to think
linearly; start with a general paragraph and work towards more and more specificity. For
example, start with a general paragraph (e.g. general questions regarding galls, effects on hosts,
biology of enemy/victim interactions), followed by a second one that introduces your specific
system (Solidago and Eurosta and its enemies). Consider starting the 3rd paragraph with … “The
purpose of this experiment is to …..”.
Each of these paragraphs must contain authoritative references.
Materials and Methods
The purpose of the methods is to inform the reader of important aspects of techniques, animals,
equipment, and conditions such that the reader could repeat the experiment. You can skip trivial
details, but you must include relevant aspects of methodology. In those cases where the
experiment was set up for you, you will have to refer back to your notes or use Goater (personal
communication). You can use subheadings if they make the section clearer (e.g. study site,
collection methods, experimental design, analyses). The last paragraph should contain a section
that describes the types of ‘analyses’ you used (correlations, ANOVA etc.).
Results
The results section should describe to the reader what was discovered in the experiment and what
you believe are the most important points. There should be two components: the presentation of
data (with tables or figures) and corresponding text that focuses the reader on the main
discoveries revealed by the data.
12
Data that do not warrant a table or figure can be reported with the text component of the results.
When data are presented in a tabular or graphical format, each is numbered sequentially, starting
with Table 1 and Fig. 1. By convention, the figures and tables should be covered in the text
component in sequential order (i.e. cover the results from fig. 1 before fig. 2 and Table 2).
Tables are accompanied by a title placed above the table that briefly and clearly describes the
contents. Each figure should be accompanied by a title placed below it. The text component
MUST contain at least one reference to each table and figure. The idea is that if you are putting
data in table or figure format, their associated results warrant specific comment. Thus, although
the results in a figure might be obvious (to you), they may not be obvious to your reader. You
will have a tendency to write a sentence something like “Table 1 provides results that show
spatial variation in gall rates on Solidago”. Fight this urge! Instead, introduce the data in the
table by a sentence such as “Gall rates were significantly higher in Population 1 than Population
2 (Table 1).”
Data should be presented only once within the results, either in a table, a figure or in the text.
Where you describe the data is up to you. Raw data are rarely given in the Results. Instead, you
should report means (always with some indication of error) and sample sizes. Statistical
treatment of the data is reported in the Results. Whenever test statistics are reported (see below),
the statistic (t, F, r), degrees of freedom, probability, and conclusion (accept or reject null
hypothesis) must be provided. Do not include the step-by-step calculations.
Gone are the days when a ruler and graph paper are used. There are now plenty of graphics
programs on the market, such as Cricket Graph and Sigma Plot. Your stats programs usually
also come with crude graphics (e.g. JumpIn, SuperAnova, StatView) and these are sufficient for
the purposes of your reports. I suggest you do not bother with Excel, other than as a worksheet.
Its’ stats and graphics sub-programs are difficult to work with if you have no prior experience
with them.
Tables and Figures should be on separate pages, and should follow the Literature Cited section at
the back of the manuscript.
Discussion
The purposes of the Discussion are to 1) interpret the data in context of other similar studies, 2)
speculate on the meaning and validity of the results, and 3) draw conclusions about the biological
phenomenon under study. Other than purely descriptive studies, most biological experiments are
conducted with a hypothesis in mind. If the data do not support the hypothesis, you should
account for the discrepancy. Possibilities include poor experimental technique, to discovery of a
previously unknown phenomenon. If discussion of the results raises additional questions, the
writer can briefly propose additional, follow-up experiments or refinements to the current
experiment.
As for the Introduction, it is useful when first starting out to arrange your thoughts in terms of
paragraphs. It is in the discussion where it is very useful to work from a well-considered outline.
If you are truly struggling with ‘writers block’, aim to match every paragraph in the Results
13
section, with a paragraph in the Discussion. It is unlikely that your discussion will stay this way,
but it is often a good place to start.
Focus on the first paragraph of your discussion. It is here where you should cover the main
results from the whole experiment, and place your results in context of other studies. The last
sentence of this paragraph often summarizes the most significant part of the work. Subsequent
paragraphs might start with ‘An alternative explanation for these results, as suggested by Smith
(1990) is …..’. It is towards the end of the discussion when editors usually allow some freedom
with interpretation (i.e. speculation). If you have difficulty leaving your ‘essay’ routes behind,
this is the only place where you might resurrect those tendencies.
Literature Cited
The key here is to choose a conventional and appropriate style, and be consistent throughout the
paper. Check a style used by one of the authors in the papers on reserve, then stick to it. Before
you hand in your paper, read over those last two sentences. Although learning where and
when to use references is a practiced skill, there is no excuse for improper format. You should
know as well that there is no answer to the question ‘how many references do I need?’. In
general, the types of reports that we deal with in my 3-4 yr classes tend to require between 10-25
references. However, I emphasize that this is a generalization and depends greatly on the type of
experiment or study that you are doing.
Extra tips
The advice described above will provide you with the necessary skills to complete a solid
manuscript. But there are still some extra tips that you might consider in order to shift your paper
from ‘good’ to ‘great’.
1. A good reviewer will decide within the first 1-2 sentences of your introduction whether you
are truly writing as an authority on a particular topic. They decide this by the content of your
sentences and by your judicious use of references. In addition, reviewers can very quickly
come to a conclusion on whether or not you really understand your data, and its’
significance. Mastering these skills is the ‘art’ of scientific writing. You have to be seen as
an authority (thus, you must know the literature cold), you have to write clearly, and you
have to convince your peers that you have something to contribute. Your first step in this
process is practice. The second is to spend some real time thinking about your data and
about the organization and preparation of your paper. It is in this fuzzy area where most
students lose their marks. Frankly, it is usually painfully easy to separate those students who
have spent time thinking about their study, and their paper, from those who have put it
together the night before the deadline.
2. Write to illuminate your information, not to impress. This is self-explanatory. It is another
way of saying ‘Too much fertilizer burns the petunias’. Be as simple as you can. My PhD
supervisor used to pretend that his 14 yr old son was sitting in the room with him, as he wrote
manuscripts, and he would first explain a difficult concept to him, then write it down. If you
are worried that you are being too flowery and verbose, go over your paper and remove every
single adverb. You will probably find that most are not needed. The same could probably be
said for many of your adjectives.
14
3. Make a statement, then back it up. This is a good habit to get into. You can back up your
statements with a good reference, or with an example, or both. ‘Not all arthropods have
reduced population sizes on heavily grazed short-grass prairie. For example, many species of
grasshopper attain maximum population sizes on heavily grazed sites (Smith, 2003)’.
4. Don’t plagiarize. I know that you will respond to this by ‘Of course not’. But be sure you
know what this statement means. DO NOT PLAGIARIZE. I have had to deal with too
many instances of plagiarism in my upper-level courses. It is an humiliating offense to deal
with. As you know, it can get you removed from the class, and even the University.
Remember that quotations are used very rarely in Scientific writing, so virtually everything
has to be in your own words.
5. Allow time for revision. This is hard to do, because we are all under time constraints. But
you will be amazed at how taking a break from your paper, then doing a revision (really a revision!), will improve your paper.
15
ORAL PRESENTATIONS
Oral presentations are developed in much the same way as that described above for written
presentations. The key difference between the two styles is that for oral presentations, you only
have a single chance to get your point across. So, it is essential that the presentation be
organized, stripped of extra details, and clear.
Time
10 min with an additional 5 min for questions. You don’t necessarily have to fill the time slot,
but you won’t be allowed to exceed it.
Format
The information in your written manuscripts will be too detailed to be covered in 15 min. I
recommend that you decide early what key points (maybe only one or two) you want to get
across, and then focus on developing these points clearly, logically, and with justification. If in
doubt, be assured that you are better off explaining one idea in sufficient detail than two
superficially.
In general, you want to present an opening that leads quickly and directly to your subject, catches
the audience’s attention and arouses their interest. They must be able to grasp as quickly as
possible the general problem (or topic) in Field Biology, on which you are focusing. This is a
good time to indicate the scope of the problem, while stating that your talk will be restricted to
its essential components. To achieve this, it is often useful to present a text-only slide that lists
the main points covered in your review, but also indicates which ones you will focus on in your
talk.
The middle portion (between 30-50% of your allotted time) should be devoted to the meat of the
issue you have introduced above (i.e. results). How you do this obviously will depend on the
nature of your topic. If you are talking about something controversial, it is here where you must
clearly present both sides of the issue. Diagrams which schematically present both sides are
useful. If your topic is focussed on something where the evidence is experimental, it is here
where you want to present diagrams on Experimental Design, flow charts of methods, and/or
clear results. Such diagrams are of tremendous use, both in helping the audience follow, and in
helping you remember what you had planned to say next.
Now you are ready to conclude, and at a critical stage. If you have been clear in the intro,
convincing in the middle part, then this section should be straightforward. Things can get
painful here, especially in cases where the audience is still unclear of the question and where the
middle section is not convincing. It is here where it is a good idea to use a text-slide to conclude
your main points.
Do not just paraphrase from your written manuscript. There will be far too much detail in the
script, especially in the Methods section. Even the Figures and Tables will probably need to be
re-done for the oral presentation. Also, as you present each component of your study, draw
conclusions as you go. Thus, one format is to present a question, then the results, followed by
the next question, and so on.
16
Visual Aids
Visuals are central to any oral presentation. The key is simplicity and clarity. All figures should
be large and easy to read. Schematic figures are usually better than those copied out of an atlas,
off a map, or out of your written script. I expect you to use a computer and laser printer to
produce the text of your visual aids. When copying figures from journals, they will almost
certainly have to be enlarged on the photocopier. Very few, if any, tables are suitable for
photocopying as they appear in the journal (they almost always contain too much information).
You must clearly indicate the source of your figures and tables, unless they are your own.
Text slides should never exceed 30 words; tables should not contain more than 6 or so cells.
Preparation
The secret behind a good talk is preparation and practice. Present your talk to a friend, or to a
mirror, both to become familiar with the sequence and to assure yourself that you don’t have to
rush to fit into the alotted time slot. Don’t think of the presentation as ‘making a speech’, but as
talking to colleagues about something that you are interested in.
Opinions vary about the use of written notes for orals. If you elect to use notes, don’t prepare a
full-blown script. Instead prepare short phrases, in a large font, so you can scan each one at a
single glance. Sometimes it is a good idea to underline words that you want to emphasize, or
places where you might speak louder or softer. Finally, practice with your notes, to make sure
they work. On the other hand, provided you are familiar with the sequence of your visuals, they
can also serve as the only prompters you will need. By relying on visuals, you will find it much
easier to maintain contact with your audience.
Delivery
Nervousness happens to everyone! It’s best to be prepared, not only for the inevitable
nervousness, but also with measures to help you relax (deep breaths, stamp your feet).
Throughout your talk, try to maintain eye contact. If you don’t maintain a relationship with your
audience, you will lose them. When using overheads or slides, face the audience and stand to one
side of the screen so you aren’t blocking the view. Face the audience, except when pointing to
the screen.
Evaluations
I will give approximately equal weighting to both content and presentation style.
17
THE USE OF SIMPLE STATISTICS IN FIELD BIOLOGY
Presented here are examples of 3 commonly used statistical tests. Each is appropriate to the types
of field studies that we usually encounter in this class. You should view these examples as
‘refresher material’ from your first-yr statistics class. I assume that you have this background
upon entering Field Biology. Further details can be found in any introductory statistics textbook.
Further examples that take a ‘Field Biology’ perspective, can be found in Fowler et al., (1998;
Practical Statistics for Field Biology), which is available upon request.
One-way ANOVA
One-way ANOVA is used to determine if there is a significant difference between two or more
samples. ANOVA focuses on the variability of the data to determine if there is a significant
difference between the means of the groups being compared (i.e., is the difference between the
group means greater than what would be expected by chance alone?). In an ANOVA the
variance is partitioned into two components: the within treatment variance (error) and the
between treatment variance, that together add up to the total variance within the overall data set.
Comparing the variability of these two components will indicate if the treatment means account
for a significant amount of the variability in the dataset. The following is an example comparing
the effect of three different herbicides on weed growth. The data and all calculations are
presented on the ANOVA worksheet.
1. Data entry and plotting
The first step is to enter the data and to plot the data. The data set we will work with has 3
treatments with 10 replicates in each. The treatments are two different herbicides used to control
weeds, plus a non-treated control. We want to test whether there is a difference in weed growth
depending on the herbicide applied. The data represent the number of plants remaining in each
plot after the herbicides have been applied. All plots had the same number of plants initially.
The next step is to plot the data to determine if there are any outliers. The data are plotted in
Figure 1. The plot of the data shows no obvious outliers. By plotting the data we notice that the
application of herbicide B resulted in the lowest number of weeds remaining followed by
herbicide A and then C. However, there is overlap in the number of plants remaining between
the herbicide applications. ANOVA will indicate whether the mean number of weeds remaining
after herbicide application is significantly different between herbicide treatments.
18
20
Response
15
10
5
0
A
B
C
Treatment
Figure 1. Plot of the data for the ANOVA example. Points represent the number of
plants remaining in each plot after herbicide application.
2. Calculating means
The next step is to calculate the treatment means and the grand mean. These calculations can be
easily done using any spreadsheet program or calculator.
3. Calculating sum of squares and degrees of freedom
Once the means have been calculated, the variability of the data can be determined. The
variability is partitioned into 2 components: between and within (error) treatment variance.
Partitioning the variance is done using the sum of squares (SS) for each of the components. The
‘Sum of squares’ quantifies the variability of the data. SS is the sum of the deviation of the data
around the mean. Three SS’s are needed to complete an ANOVA. SST = total sum of squares, a
measure of the total variability of the data around the grand mean. SSE = error sum of squares, a
measure of variability of the data within each treatment around the treatment mean. SSB =
between sum of squares, a measure of the variability of the treatment means around the grand
mean. These three measures of variability have a simple relationship:
SST = SSE + SSB
Sum of squares are dependent on the number of observation in the data set. The greater the
number of observations, the greater the sum of squares. Therefore, we need to compare sum of
squares per the number of degree of freedom associated with each SS (i.e., the variance).
Degrees of freedom (d.f.) are the number of independent pieces of information contributing to
the statistic, in this case the variance. For example, if nine out of the ten deviations from the
mean are known, the tenths one is predetermined because all the deviations must sum to zero.
Therefore, in the herbicide example, the d.f. associated with the SST is 30-1 or n-1, where n is
the number of observations. The d.f. for the SSB is 3 –1 or k-1 where k is the number of groups
and for SSE is 30-3 or n-k where is n is the total number of observations and k is the number of
groups or treatments.
19
4. Calculating mean squares
Dividing the sum of squares by the appropriate d.f. results in the variability per degree of
freedom, called the mean square (MS). Therefore, in our example, the MS associated with
herbicide is
MSH = SSB / 2,
the variance between plots given different herbicides. The MS associated with the SSE, with
treatment variance is
MSE = SSE / 27,
The variance within plots given the same herbicide. Finally the MS associated with the total
variance is
MStot = SStot / 29.
5. Calculating the F-ratio
If the herbicides have no effect on the weed seedlings, than it would be expected that the MS for
the herbicides (MSH) would equal the MS of the error (MSB / MSE =1). This ratio is termed the
F-ratio in an ANOVA.
F = MSH / MSE
A table of critical F values is used to determine if the F value computed is significantly greater
than 1. The F calculated in our example is 20.77 which is greater than the critical value of F
with 27 and 2 degrees of freedom. Therefore, the null hypothesis can be rejected and we can
conclude that there is a significant effect of herbicide on weed growth.
20
ANOVA Worksheet
Treatment
1.
2.
Mean
A
B
5
12
9
8
11
12
7
14
6
8
9.20
1
2
7
3
8
9
4
2
6
3
4.50
C
10
13
16
12
17
11
18
9
15
10 Grand
8.93
13.1
Herbicide
Data
(Y)
Treatment
Means (T)
Grand
Mean (G)
(Y-G)
A
5
9.2
8.93
15.44
17.64
0.07
A
A
A
A
A
A
A
A
A
B
B
B
B
B
B
B
B
B
B
C
C
C
C
C
C
C
C
C
C
12
9
8
11
12
7
14
6
8
1
2
7
3
8
9
4
2
6
3
10
13
16
12
17
11
18
9
15
10
9.2
9.2
9.2
9.2
9.2
9.2
9.2
9.2
9.2
4.5
4.5
4.5
4.5
4.5
4.5
4.5
4.5
4.5
4.5
13.1
13.1
13.1
13.1
13.1
13.1
13.1
13.1
13.1
13.1
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
8.93
9.42
0.00
0.86
4.28
9.42
3.72
25.70
8.58
0.86
62.88
48.02
3.72
35.16
0.86
0.00
24.30
48.02
8.58
35.16
1.14
16.56
49.98
9.42
65.12
4.28
82.26
0.00
36.84
1.14
7.84
0.04
1.44
3.24
7.84
4.84
23.04
10.24
1.44
12.25
6.25
6.25
2.25
12.25
20.25
0.25
6.25
2.25
2.25
9.61
0.01
8.41
1.21
15.21
4.41
24.01
16.81
3.61
9.61
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
0.07
19.62
19.62
19.62
19.62
19.62
19.62
19.62
19.62
19.62
19.62
17.39
17.39
17.39
17.39
17.39
17.39
17.39
17.39
17.39
17.39
3.
Sum of Squares (SS)
d.f.
2
(Y-T)
2
Total
Error
611.87
29
241.00
27
(T-G)
2
Be tw ee n
370.87
2
4. And 5.
Source of
Variation
Herbicide
Error
Total
Sum of
Squares
370.87
241
611.87
d.f.
2
27
29
Mean
Square
185.435
8.92593
F
20.7749
21
Simple Linear Regression
Regression looks for a linear relationship between two variables. The two variables involved are
called the dependent variable (Y), and the independent variable (X). The regression equation
fitted to the data, formulated from the analysis, describes the relationship between the
independent and dependent variable. The general form of this equation is:
Y = a + bX
where: a is the y-intercept and b is the slope of the regression line. The regression equation is the
line for which the sum of the squared vertical deviations between the data points and the
regression line is minimized (i.e., ∑(Y-Ybar)2 is minimized). This can be visualized by pivoting
the regression line on the Xbar, Ybar coordinate. The best fit is when the sum of the deviations
between the data points and the regression line is minimized (figure2). The variability around
the regression line is the error sum of squares (SSE).
16
15
Foraging Interval (min)
14
13
12
11
10
9
8
7
6
5
60
70
80
90 100 110 120
Volume Consumed (µL)
130
140
150
Fit Mean
Linear Fit
Figure 2. Plot of the data showing the mean and regression line. The regression line crosses the
mean line at the Xbar, Ybar coordinate. The best fit for the regression line is when the sum of
the deviations between data points and the regression line is minimized.
Regression analysis is used to describe the association between the X and Y variable. Is
the relationship positive or negative? As well, regression can be used to predict Y given a
certain value for X once the equation has been determined. The R-squared statistic is often
calculated when doing regression analysis. The R-squared value is the percentage of the
variability in the dependent variable that is explained by the independent variable. The
following example will demonstrate how to complete a regression analysis, calculating both the
regression equation and the R-squared value. All calculations are demonstrated on the
Regression work sheet.
The data used in this example are observational data exploring the relationship between
nectar consumption and feeding interval of hummingbirds. We are interested in describing the
relationship between nectar consumption and foraging bout interval. Regression analysis allows
22
the investigator to determine the detail of the relationship between nectar consumption
(independent variable) and foraging bout interval (dependent variable).
1.Data entry and plotting
Enter the data into a spreadsheet in XY pairs. That is, enter the X values in one column and the
corresponding Y values in the next column (see regression worksheet). The next step is to plot
the data and examine the data for outliers. The general trend of the data can also be determined
by plotting the data (Figure 2). The plot of the data reveals that there is a positive relationship
between nectar consumption and foraging bout interval. Regression analysis will determine if
this relationship is statistically significant.
2. Calculating variable means
Once the data have been entered, the means of both the independent variable and the dependent
variable need to be calculated.
3. Calculating the regression equation
Several values are needed to formulate the regression equation. The slope (b) of the regression
line is found using the following formula:
b = ∑XY- (∑X)(∑Y)/n
∑X2 – (∑X)2 / n
The values used in calculating the slope of the regression line can be easily calculated using a
spreadsheet program (see regression worksheet). The slope for the regression line between nectar
consumption and foraging interval is 0.073.
4. Calculating the y-intercept
The y-intercept is calculated using the following formula:
a = bXbar – Ybar
where a is the y-intercept, b is the slope of the regression line and Xbar and Ybar are the means
the X and Y variables, respectively. The y-intercept for our example is 3.84.
5. Calculating sum of squares
The sum of the deviations of the data from the grand mean for the Y variable is the SST. SST is
the sum of the differences between the Y values and the mean of Y (Ybar). The SSR is the
deviation between the regression line and the mean line. The formula to calculate the SSR is
SSR = b(∑XY - ((∑X) (∑Y)) / n)
where b is the slope of the regression line, ∑XY is the sum of the X variable times the Y
variable, ∑X is the sum of X variable, ∑Y is the sum of the Y variable, and n is the number of
paired observations. The sum of squares error (SSE) is the variability around the regression line.
This is the variability not explained by the regression line. SSE is calculated with the following
formula
SSE = SST – SSR.
6. Calculating the F-ratio
The SSR, SSE, and the appropriate degrees of freedom are used in calculating the F-ratio. Mean
squares are calculated by dividing the sum of squares by there appropriate degrees of freedom.
23
These calculations for the hummingbird example are given in the ANOVA table on the
regression worksheet. The F-ratio is calculated using the following formula
F = MSR
MSE
A table of F values is used to determine if the regression line accounts for a significant amount of
the variability in the data. F for our example is 9.85, which is greater than the critical value of F
(5.32) with 1 and 8 d.f. Therefore, we can conclude that there is a relationship between nectar
consumption and foraging bout interval. The greater the volume of nectar consumed the longer
the foraging bout interval.
7. Calculating the R2 statistic.
The R2 statistic is a ratio of the variability in the data explained by the regression line (SSR) to
the total variability in the data (SST) . R2 is calculated using the following formula
R2 = SSR
SST
24
The R2 for our example is 0.55. Therefore 55% of the variability in foraging bout interval is
explained by volume of nectar consumed.
Regression Worksheet
1.
X variable
Y variable
Volume
Foraging
2
Consumed
Interval
X
(µL)
(min.)
90
11
8100
65
6
4225
90
13
8100
85
12
7225
90
11
8100
115
14
13225
110
13
12100
80
9
6400
130
11
16900
140
15
19600
2.
•
995
115 103975
Means
99.5
12
(Xbar )
(Ybar )
2
XY
Y
121
36
169
144
121
196
169
81
121
225
990
390
1170
1020
990
1610
1430
720
1430
2100
1383
11850
(Y - Y bar )
(Y - Y bar )
-1
-6
2
1
-1
3
2
-3
-1
4
5.
0
2
0
30
2
0
0
6
2
6
0
12
SST
60.50
3.
b=
•Y- (•
X
X)(•
Y)/
n
•
X 2 – (•
X) 2 / n
5.
SSR = b(•
XY - ((•
X) (•
Y)) / n)
SSR = 0.082(11850 - ((995)(115) / 10)
SSR =
b=
33.42
11850 - (995)(115)/10
103975 - (95
2
) / 10
SST= SSR +SSE
b= 0.082
SSE= SST-SSR
SSE= 60.5 - 33.42
SSE= 27.08
4.
a= Ybar + bXbar
a= 12 + 0.082*99.5
a= 3.84
6.
F=
F-Ratio
Source of
Variation
Model
Error
Total
Sum of
Squares
33.42
27.08
60.5
Mean Square
(SS / d.f.)
1 (MSR) 33.42
8 (MSE) 3.39
9
d.f.
MSR
MSE
F
9.858
7.
SSR
2
R = SST
33.42
2
R = 60.5
2
R = 0.55
25
t-test
A t-test is used to determine if two sample means (µ1, µ2) are different. The null hypothesis is
that µ1 - µ2 = 0. To test the alternative hypothesis that µ1 - µ2 ≠ 0, the t statistic is calculated.
The formula to calculate t is:
t =
(Ybar1 – Ybar2) – (µ1 - µ2)
√((s12 + s22) / n)
where Ybar1 and Ybar2 are the means of the two groups being compare, s12 and s22 are the
variance of the two groups, and µ1 and µ2 are the sample means. We are testing the null
hypothesis that µ1 - µ2 = 0, therefore (µ1 - µ2) in the formula can be replaced with 0.
Assumptions of the t-test are: data are continuous and normally distributed, and the variance of
the two sets is homogenous.
We will work through an example of a t-test by comparing the pronotum (head) width of male
and female crickets. Ten crickets of both sexes were selected randomly from population of
crickets collected in pitfall traps.
The data are entered into two separate columns: female and male. Next, the mean and the
variance are calculated (see t-test worksheet) for both sexes. The mean and variance of each
group are the only values needed to determine the t- statistic. Substitute the mean and the
variance for both groups into the formula to derive the t-statistic.
To determine if the difference is significant, the degrees of freedom need to be calculated.
Degrees of freedom are found using the following formula;
d.f. = 2(n-1).
Consult a table of critical values of the t-distribution to determine if the t-statistic is significant at
the appropriate degrees of freedom. For the cricket example the observed value of t (1.879), is
les than the critical value of t with 18 degrees of freedom (2.101). Therefore, the pronotum
width of male and female crickets is not significantly different. We fail to reject the null
hypothesis.
T-test worksheet
Y1
Y2
Female
(Y 1 bar - Y 2 bar) - (µ
Male
5.3
5.04
5.35
5.57
4.69
4.93
5.46
5.02
5.29
4.99
5.26
4.85
4.24
5
5.03
5.2
5.09
5.23
4.92
4.96
Mean
5.19
4.96
Variance
0.07
0.08
10
10
Count
t=
¦((s
1
+s
2
2
1
- µ 2)
) / n)
(5.19 - 4.67) - (0 - 0)
¦((0.07 + .08) / 10)
t=
t=
2
1.8790002
d.f.= 2(n-1)
d.f.= 2(10-1)
d.f.= 18
t 0.05(18)
= 2.101
t critical
26
1.
2.
3.
4.
APPENDIX
Table of Random Numbers…………………………………………………………………
Distribution of parent and hybrid cottonwoods in Southern Alberta ……………………
Natural history of the Goldenrod ball gall – the key players ………………………………
Leaf morphology of parent and hybrid cottonwoods ………………………………………
28
29
30
31
27
28
29
30
31