GERMINATION CUES FOR SEEDS
Order Description
PRACTICAL 2A GERMINATION CUES FOR SEEDS
AIMS AND OBJECTIVES
At the end of this Practical, students should be able to:
� frame a question that can be answered using scientific method;
� design an experiment to answer a scientific question;
� execute the experiment;
� analyse the results;
� present the findings as a talk with PowerPoint slides.
INTRODUCTION
In many vegetation types both here and overseas, fire is a regular occurrence. It is common to find that germination of seeds of plant species present in these vegetation types is limited to the months immediately following a fire, with little or no germination in the inter-fire period. This flush of germination occurs when there is plenty of light at the soil surface; nutrient levels in the soil are usually elevated, as the fire has removed the previous vegetation cover, and nutrients present in the biomass are returned to the soil as ash. Further advantages to germinating in the immediate post-fire period may include reduced activity of soil-borne pathogens (killed by the fire), and satiation of seed/seedling predators with overwhelming numbers of seeds/seedling.
Observation of the flush of germination that follows a fire then leads to the following question: how do the seeds �sense� that a fire has occurred? There are a number of signals or �cues� that might operate to initiate germination: a partial list would include
� the pulse of heat (temperature achieved, for given time)
� chemicals released by the burning process
� elevated nutrient levels in the soil
� elevated light levels at soil surface
In order for us to know which one or ones of these (or other) factors or �cues� might stimulate germination, we need to conduct experiments to test the effect of each factor (or combination of factors) on germination.
For one group of plants something is known about how the seeds �sense� a fire; this group is the members of the Pea Family (Fam. Fabaceae) and the related Wattle Family (Fam. Mimosaceae). The seeds of members of these two plant Families will not germinate readily when first released from the parent plant – they are said to be dormant (Auld & O�Connell 1991, Vleeshouwers et al. 1995). In this case, the dormancy is imposed by a �hard� seed coat which prevents the seed taking up water, the critical first step in germination (Langkamp 1987, Morrison et al. 1992, 1998). Heating (or scarifying) the seed breaks this mechanical barrier to water uptake, and allows germination to proceed (Burrows et al. 2009).
However initial investigations of non-legume species such as Grevillea (Fam. Proteaceae) showed that heat shock (by itself) had little effect on germination (Auld & Tozer 1995).
Interest recently has centred on smoke as a source of a germination signal. A positive germination response to smoke or smoke-derived extracts has been detected in 12 fynbos species from South Africa, including species in the Asteraceae, Ericaceae, Restionaceae and Proteaceae (Brown 1993, Brown et al. 1993), Themeda triandra from South Africa (Baxter et al. 1994), Nicotiana attenuata from America (Baldwin et al. 1994), and 45 species of Autralian native plants from West Australia, including species in the Rutaceae, Dilleniaceae, Proteaceae, Myrtaceae and Cupressaceae (Dixon et al. 1995). This range of species from different continents suggests that stimulation of germination by smoke or smoke-derived compounds may be common in fire-prone vegetation (Baxter et al. 1994). The response of Themeda australis to smoke was investigated by Clarke and French (2005).
In some cases, the two fire cues combine to affect germination: this combination can be positive e.g. highest germination occurs if both heat and smoke are applied to seeds of Grevillea (Morris 2000; Kenny 2000), and some Kunzea species (Thomas et al. 2003). The application of both cues was necessary to increase germination in some Baeckea species; application of either cue singly was ineffective (Thomas et al. 2003). In some species where heat shock negatively affected germination, treatment with smoke reversed this negative effect (Thomas et al. 2007).
To investigate these phenomena we will use seeds from two different native species (TBC) and an agricultural cereal species for comparison.
EXPERIMENTAL DESIGN
Rather than have the whole class follow the one experimental design, each group will discuss amongst themselves, and design and execute one experiment. For the conduct of the experiment, we have provided the following resources:
� seeds of two native species and a cereal;
� ovens set to cover a range of temperatures (25oC, 80oC, 100oC and 120oC) (heating effects);
� smoked water and smoked vermiculite (smoke cues);
� compost, pots and incubators for germination.
Some of the questions your experiment might ask are:
1. What is the best temperature of �heat shock� to stimulate germination (for given duration of heat exposure)?
2. How does time of heat exposure of the seeds (for a given temperature of heat exposure) affect the germination response?
3. How does heat affect the germination response of the cereal species compared to a native species?
4. Do both native species respond in the same way to the same fire-related germination cues?
5. Does a species show a germination response to smoke, or to smoke + heat?
Whatever question from the above list (or any other reasonable question you pose), you will have to think about the requirements of good experimental design:
� your treatments
� your controls
� numbers of replicates.
READ BOX 1 BEFORE YOU DESIGN YOUR EXPERIMENT.
BOX 1 EXPERIMENTAL DESIGN
In order to answer the question or questions that you ask in this Practical, you will have to design an experiment. A brief word about some standard experimental designs, and the methods used to analyse them, is appropriate here. As we have already seen in this Subject, there is a close link between the way data is gathered, and how it can be analysed: if inferential statistics are to be used in analysis, the data must be obtained in a way that satisfies the assumptions of these tests.
One-way or Single Factor Design
Here, different levels of one factor are applied to the experimental subjects, and their response is measured in some way.
In the context of the seed germination experiment, an example of a Single Factor design would be to vary time of exposure of wattle seed to a given, fixed temperature (say 800C).
Different levels of the Time of heat exposure factor are applied to batches of seeds � e.g. 1,2,5,10 and 20 minutes. The data from such an experiment might look like Table 2.1
The results of such an experiment would be analysed by a One-Way Analysis of Variance (see Banksia Practical).
Table 2.1 Sample data from a Single Factor seed germination experiment.
TIME OF HEAT EXPOSURE 1 min 2 min 5 min 10 min 20 min
replicate 1 0 0 4 9 0
replicate 2 0 2 7 10 0
replicate 3 1 0 4 7 2
replicate 4 0 1 3 6 1
Two-way or Factorial Design
Here we add a second Factor to the experiment. If all levels of the second factor are tested with all levels of the first Factor, the design is said to be a Factorial Design.
Let�s say that in addition to testing Time of heat exposure of the seed as a Factor, you wished to add Temperature of exposure as the second factor – say 600C, 800C, 1000C and 1200C. The combination of the two Factors can be seen in Table 5.2:
Table 5.2: Factorial combination of two experimental factors (1. time of heat exposure; 2. temperature of heat exposure). Every Temperature of exposure is tested at each Time of exposure.
TIME OF HEAT EXPOSURE 1 min 2 min 5 min 10 min 20 min
TEMP. OF HEAT EXPOSURE 600C 600C 600C 600C 600C
800C 800C 800C 800C 800C
1000C 1000C 1000C 1000C 1000C
1200C 1200C 1200C 1200C 1200C
In this case, replicate petri dishes would be set up as above, and the number of seeds germinated in each one scored at the end of the observation period.
These data would then be analysed by a Two-Way Factorial ANOVA, which would let you judge the significance of each of the two Factors used (Time of exposure, Temperature of exposure), and any interaction between the two Factors (if one exists).
To do a Two-Way ANOVA, and interpret it, see Appendix IV and your Demonstrator.
Replication, Randomness and Independence in experimental design
We have already met the requirements for data to be random and independent when we have used t-tests and ANOVA in earlier Practicals. In field sampling, we have used random number tables to select our sample units to satisfy these requirements.
In experiments, there are other steps to take to ensure that the measurements we take on our experimental subjects are replicated, random and independent.
Replication: the treatment is applied not merely to a single experimental subject, but to a number of them, and the average response of subjects to the treatment is calculated. The response of a single subject to an experimental treatment, while interesting, cannnot be taken as indicative as the general response of subjects to the treatment. Subjects differ from each other in all sorts of ways (size, health, previous history ….) and all of these affect the response of the subjects to the treatment. What we are trying to see is whether there is any general response of subjects to the treatment, given this background level of variability in the population.
Randomness: this is required in experiments so that any effects of treatments we detect can be clearly inferred to be the result of the treatment, and not some artefact of the way we have run the experiment. Steps which we take to obtain random data, and so avoid getting a spurious result, would include
� random allocation of subjects into treatments. If for example in our seed experiment, we put all the big, healthy-looking seeds into one treatment, and all the small, unhealthy-looking seeds into another treatment, any differences between the treatments could simply reflect this initial non-random allocation of seeds into treatments. So subjects must be randomly allocated into treatments to avoid bias in the estimation of treatment effects.
� random arrangement of pots after treatment. The pots need to be left for several weeks, to allow germination to occur. It is important to randomly arrange the pots in a stack to be left in the growth cabinet – if all the replicates of one treatment are put together, and there is some peculiarity of the growth cabinet where they sit (warmer, colder, ligher, darker than the rest of the cabinet) germination of the seeds in these replicates may be affected by the particular growth conditions in that part of the cabinet. This would be confused with a treatment effect. The requirement for random arrangement of experimental subjects in a growth area is referred to as random interspersion. So random interspersion of the experimental subjects in the incubation area guards against spurious effects due to peculiarities of the growth cabinet.
Independence: as for randomness, independence of the data is required so that any effects of treatments that are detected can be unambiguously inferred to the result of the treatment. Many of the steps taken to ensure randomness will also ensure independence, but there are some additional procedures involved for indepencence of data from experiments:
� independent application of the treatment to replicates. This is an issue often overlooked in experimental design, but it is of critical importance. We have already seen that replication of subjects is required, to estimate the average response of subjects to an experimental treatment, against the background of variability due to other sources present in the population. Given that replication of subjects is required, replication of applications of the treatment are also required for the same reason – a single application of the treatment is not adequate to estimate the general effect of the treatment on the population. If only a single application of the treatment is made, even if to replicate subjects, any peculiarity of that single application would affect all the subjects to which is was applied, and this could be confused with a treatment effect. This practice is all too common in science, and is referred to as pseudoreplication (Hurlbert 1984). Thus, there must be repeated and independent applications of any experimental treatment to experimental subjects. See the discussion of this problem in seed germination experiments in Morrison & Morris (2000). A true replicate is thus the smallest experimental unit to which a treatment is independently applied. If the treatment is independently applied to replicate experimental units, the deviations of the responses of these replicates from the treatment mean are independent of each other, and are independent estimates of the background variability (or �error�) in the system. If the treatment is only applied once, to replicate subjects, the deviations of the responses of these replicates from the treatment mean are not independent of each other – they are linked by being part of the single application of the treatment.
Controls
In order to tell whether a given treatment has had an effect, we compare the response of subjects receiving the treatment, to the same response of subjects not receiving the treatment. Subjects not receiving the treatment are called controls. Sometimes, when applying the treatment involves a number of steps, we will use procedural controls (handled in same way as treated subjects, without receiving the actual treatment. Thus in trials of a new drug, control subjects receive a dummy pill, called the placebo, to imitate the effects of taking a substance on the subject). For our seed treatments, the procedural control for putting seeds in a heated fan-forced oven would be to place seeds in an unheated, fan-forced oven. If a significant germination response is detected in the heat-treated seeds, it can be safely concluded that the heat produced the effect (and not the convection oven by itself).
METHOD
Once you have designed an experiment, talk over the details with Natalie and/or Alison. After the discussion, execute the experiment (as originally designed, or modified after discussion) using the methods below.
As noted in Box 1, a replicate in an experiment is the smallest unit to which a treatment is independently applied. In our case, treatments will be applied separately to a seed lot of 6 – 10 seeds in a pot, and the response measured as the percent germination of the seeds at the end of the experiment. Thus the replicate is the seed lot (and not the individual seeds in it).
To satisfy the requirement for independent application of smoke and heat, seed lots must be put treated one after another (and not all put in simultaneously). This ensures repeated, independent application of the treatment. .
Treating seed samples with heat
1. place an individual seed lot in a glass petri dish, and leave it in the oven (with the lid off) for specified period of time.
2. Repeat for your remaining replicates.
Treating seed samples with smoked water or smoked vermiculite
Choose to either regularly water your seeds with a dilution of smoked water or use the smoked vermiculite as top dressing. Why would you choose one over the other?
Germinating seed lots
Seeds will be planted in compost in pots and watered regularly. You will be in charge of watering them. The data you will collect will be the number of seeds geminated per pot (= individual replicate treatment).
DATA COLLATION AND ANALYSIS
Calculate the
� mean no. of seeds germinating in each treatment
� Standard Error or Confidence Limits around the mean.
This is the basic step of summarizing your results.
Do not graph results for individual replicates within treatments.
Decide what type of graph best presents the results you have obtained.
Since each group performed its own experiment, it is difficult to specify in advance what will be the most appropriate method of analysis for your data.
Consult with a demonstrator to see what (if any) statistical analysis can be used for your data. This goes back to your
� original hypothesis, and the
� experimental design you used to test it.
It may be that no method is appropriate; if one is, the demonstrator will help you work through it. You may be able to use
� t-test (for comparing 2 samples)
� 1-Way ANOVA (>2 samples, single factor design)
� 2-Way ANOVA (two factors varied in design)
You may need to check Homogeneity of Variances amongst samples being compared (Banksia Practical); transform data if variances are heterogeneous.
ASSESSMENT OF SEED GERMINATION PRACTICAL
Assessment of Practical 2A �Germination of seeds� will be in the form of
� joint Group Power Point presentation (50% of marks)
� talk by each student (50% of marks)
POWER POINT PRESENTATIO
each Group will present their experimental results of the seed germination Practical as a short talk supported by Power Point slides (15 minutes� total). The presentation will have the following sections:
1. Introduction:
A clear and concise introduction covers all required information, with comprehensive context and background information, creative aims and research questions.
2. Methods:
Clearly and concisely states the study design, lists all steps in the methods and describes all methods to handle and analyse the data
3. Results:
highly effective use of table and /or graphs that highlight all the trends and patterns of data; Text gives clear, concise summary of the results
4. Discussion/Conclusions:
Answers with good insight into underlying processes are given; all are clearly and correctly linked to results. References are well used to convincingly confirm/refute/extend each of the findings, references come from a wide range of quality sources.
5-References:
Consistently correct use of the Harvard referencing style.
(I want great report and not copy and must be Australian English and easy understood please)
