2017 Planting Experiment
- Amount to Plant
2/13/2017
For the 2017 planting I
have to decide how much I will plant. First I needed to start with shelf space.
For the Portulaca experiment, I was
using one shelf for experimentation. This did keep me busy, but I wanted to do
more. I added another unit with 3 additional shelves. For this planting I will
use only the new shelves and keep my old shelve vacant for random plants. With the
number of shelves determined, 3, I can focus on what is on the shelves.
I need to decide how
many units per shelf. For the Portulaca
experiment, the shelf could hold 98 glass jars. I discussed the matter with a
colleague who suggested I use plastic cups. I evaluated the idea. Plastic cups
fit 72 per shelf. This is a 27% decrease in the number of units on a shelf. For
every 4 shelves of jars, it is as if I get a bonus shelf when compared to the
cups. Jars are about $10.00 per 12 units that would total $245.00 for 3
shelves. Cups are about $3.00 per 24 units that would total $27.00 for 3
shelves. Cups are nearly 11% of the cost of jars. Once the experiment is
complete, I will have to clean the jars. If I clean jars at the speed I have
the Portulaca experiment, it may take
a year. I’m not lazy (although some will argue), just busy. With the plastic
cups I can dump the dirt then recycle them, they are #2 plastic, within a day. Jars
take up space. I have to keep them boxed up before use and when the experiment
is completed. The plastic cups stack on each other, and although seems
wasteful, are cheap and disposable. I could clean and reuse the cups, but that
will be based on their condition and my cost/benefit ratio that won’t be
discussed. The pros for the cups out weight that of the jars (Table 1). I tried
the jars before with good results. It is worth testing cups as they are similar
but have low costs. The 3 shelves, 72 cups per shelf, in total can hold 216
cups. With cups selected, I can plan how many and what kind of seeds I will
plant.
Table 1. Pros and Cons
of Jars and Cups.
|
Jars
|
Cups
|
Units per shelf
|
27% more per shelf
|
less per shelf
|
Cost
|
$245.00
|
$27.00
|
Amount of work
|
clean when done
|
dispose recycle
|
Storage
|
must house
|
collapsible
|
|
|
|
Pro:
Color
|
|
Cups Win
|
Con:
Color
|
|
|
To start, determining
the species I will plant is based on different factors. Half depends on how
many cups do I want to appropriate to a species. The other half depends on how
many seeds I have available. Then I must combine those findings to determine
how many seeds per how many cups. A literature search can help justify numbers
and provide an idea on what type of germination rates to expect. Rather than no
previous background, I have decided a non-intensive literature search to cite a
few germination percentages. This will provide a base background and justification,
yet no doubt will lack the full benefit on an intense search. If germination
rates are poor, then a deeper investigation of the literature will be
warranted. Until then, the few examples I find will suffice my curiosity on the
subject.
I plan to devote one
shelf to Baptisia species that includes B. alba, B. perfoliata, and B. lanceolate. I wish only to examine
germination, growth, and compare those characteristics among species. I will
have 24 replicates (cups) for each species. I will maintain one plant per cup.
If more seeds germinate than one, the more vigorous plant will be allowed to
grow and I will remove the others.
Based on literature
about the genus germination, I must decide how many seeds I will place within
each cup. I searched for literature online and developed a table taking note of
overall percent germination for each experiment (Table 2). I then averaged the
percent germination to determine 42% of seeds germinate. I understand this is a
rough estimate as each experiment contained different conditions, yet it gives
me a baseline that I will use as justification. With the baseline 42%
germination, I can estimate how many seeds I need per cup.
Table 2. Baptisia
germination literature search.
Species
|
%
|
Citation
|
Type
|
B. arachnifera
|
90%
|
Pattavina within Estep 2011
|
Cold
|
B. australis
|
90%
|
Dirr 1987
|
Scarring
|
B. australis
|
90%
|
Boyle 2005
|
Scarring
|
B. australis
|
80%
|
Bratcher 1993
|
Cold 10 weekd
|
B. australis
|
80%
|
Bratcher 1990
|
Cold 10 weeks
|
B. australis
|
70%
|
Bratcher 1993
|
None
|
B. australis
|
70%
|
Bratcher 1990
|
No Stratification
|
B. australis
|
70%
|
Bratcher 1990
|
Cold 8 weeks
|
B. arachnifera
|
60%
|
Estep 2011
|
No Stratification
|
B. australis
|
60%
|
Bratcher 1990
|
Cold 2 weeks
|
B. australis
|
60%
|
Bratcher 1990
|
Cold 6 weeks
|
B. australis
|
45%
|
Boyle 2005
|
None
|
B. australis
|
45%
|
Boyle 2005
|
Water soak
|
B. australis
|
40%
|
Bratcher 1990
|
Cold 4 weeks
|
B. alba
|
30%
|
Ksiazek 2014
|
Ground level
|
B. tinctoria
|
22%
|
Wennerberg 2005
|
Cold 3 weeks + acid, in Field
|
B. lanceolata
|
15%
|
Simons 2010
|
Less than in all field treatments
|
B. tinctoria
|
13%
|
Wennerberg 2005
|
Cold 3 weeks + acid, in Greenhouse
|
B. lanceolata
|
11%
|
Estes 2007
|
Lab
|
B. australis
|
10%
|
Boyle 2005
|
Scarring and water soak
|
B. lanceolata
|
6%
|
Estes 2007
|
Lab
|
B. leucantha
|
2%
|
Voigt 1977
|
Cold 8 weeks moist
|
B. lanceolata
|
2%
|
Estes 2007
|
Nursery
|
B. leucantha
|
0%
|
Voigt 1977
|
None
|
B. alba
|
0%
|
Ksiazek 2014
|
on green roofs
|
Average
|
42%
|
From 25 different
experiments
|
At the time I was
disappointed in my research on the understanding of statistical power but I am
pleased to put it to use in determining the number of seeds per cup. Statistical
power is used to estimate the probability you will determine something based on
numeral factors. It can also be used to estimate how many samples you need to
increase the probability you will obtain a result. The philosophy and
application of statistical power is debatable on many levels, yet if used
wisely can provide justification to experimental design. Although many
equations for statistical power and sample size are complex, when much is
unknown rough estimate equations are developed. I find these rough estimate
equations useful in that a simpleton such as myself can use them.
The “Rule of 3s” to determine sample size is explained
(
http://www.vanbelle.org/chapters%5Cwebchapter2.pdf).
The equation divides 3 by an estimate. Through the literature search, 42%
germination is my known estimate. I just divide 3 by 0.42 then round up to
obtain 8 seeds as an estimate of a sort of guarantee to obtain 1 seed
germinating per cup (Figure 1).
x=
|
0.42
|
|
n=
|
3
|
/
|
n
|
|
|
|
n=
|
3
|
/
|
0.42
|
|
|
|
n=
|
|
|
7.142857
|
Figure 1. Spreadsheet “machine”
to calculate sample size using a percent estimate. In the blue box I input the
estimate percent. The orange box returns the suggested sample size. I obtained the percent estimate from table 2 above.
I will further explain
this witchcraft. As an example: A fisherman catches 5 fish, on average, every
hour. We divide 1 hour by 5 fish to obtain the value 0.2. The 0.2 means for
every 0.2 hours, a fish will be caught, on average. By multiplying 60 minutes
by 0.2, we can determine that every 12 minutes a fish is caught on average.
Some mad scientist came up with the idea of 3s: multiply the estimate by three
and you have a good chance to obtain the expected result. We take 12 minutes
and multiple it by 3 to obtain 36 minutes. If the fisherman fishes for 36
minutes, he has a good chance to catch a fish. If you gamble like a scientist,
then you will put your money on the 36 minute mark that the fisherman will
catch at least one fish. That was still complicated.
Example explained
another way. A fisherman catches 5 fish per hour, on average. For one hour = 5 fish.
We apply the rule of 3s equation. We divide 3 by 5, the result is 0.6. Meaning
in 0.6 hours, you have a 95% chance to catch 1 fish, statistically speaking.
Multiplying 0.6 hours by 60 minutes is 36 minutes. Between both explanations, I
think the point may get through (to myself).
Now I apply the concept
of the rule of 3s to my experiment. For every 100 seeds, 42 germinate (42%), my
on average rough estimate taken from the literature. I take 42/100 to obtain
0.42. I divide 3 by 0.42 to obtain 7.142857. Only whole seeds can be used. No
such thing exists as half a seed grows half a plant. It is one seed for one
plant. Not 2 seeds for one plant. Not 0.5 seeds for one plant. Not 0.142857
seeds for one plant. Only 1 seed for one plant. So I round my 7.142857 estimate
to 8 seeds. With an estimate of 42% germination, I need to plant at least 8
seeds per cup to obtain at least 1 seed germinating.
The rule of 3s
calculation can also work in reverse with the same formulas. If 7.142857 is
entered, then 0.42 will return. If seeds were limited, then the return on the
function would estimate what percent of germination you may obtain based on
seed availability.
The result of 8 is a
good suggestion. “In some disciplines, group sizes of eight are almost
universal and some referees may go so far as to reject a paper which deviates
from this norm,” (
http://www.ebd.csic.es/documents/240051/0/Sample_size_and_power_analysis.pdf).
The quote references 8 for the number of samples to determine a statistical
difference between groups. My application of 8 is used to obtain a statistical
guarantee of germination within a cup. Applied differently the concept is the
same. As a side note, understanding this explains that the “8” rule as having a
slightly less than 0.4 (or slightly less than 40%) assumption on the estimate.
That is the theory, but
can I apply it to actual application? 24 x 8 = 192. I have enough seeds for B. alba and B. perfoliata, but unfortunately for B. lanceolate I have 181
seeds. I will have to adjust.
With seeds, I often
like to retain some in case. In time I have a collection of seeds. Older seeds
are not as good as newer seeds due to decreased viability. For this I will get
into the mindset to use all seeds.
Although I have gone
through the process shown above, I must adjust my thoughts for the project. I
determined at least 8 seeds, but I only have enough for 7 seeds per cup. I then
return to the literature. I plan to use cold stratification and will need to
determine how many weeks I will keep the seeds in the refrigerator. I filtered
the table for cold stratified seeds in germination experiments.
Table 3. Refined Baptisia germination literature search.
Species
|
%
|
Citation
|
Type
|
B. arachnifera
|
90%
|
Pattavina within Estep 2011
|
Cold
|
B. australis
|
80%
|
Bratcher 1993
|
Cold 10 weekd
|
B. australis
|
80%
|
Bratcher 1990
|
Cold 10 weeks
|
B. australis
|
70%
|
Bratcher 1990
|
Cold 8 weeks
|
B. australis
|
60%
|
Bratcher 1990
|
Cold 2 weeks
|
B. australis
|
60%
|
Bratcher 1990
|
Cold 6 weeks
|
B. australis
|
40%
|
Bratcher 1990
|
Cold 4 weeks
|
B. tinctoria
|
22%
|
Wennerberg 2005
|
Cold 3 weeks + acid, in Field
|
B. tinctoria
|
13%
|
Wennerberg 2005
|
Cold 3 weeks + acid, in Greenhouse
|
B. leucantha
|
2%
|
Voigt 1977
|
Cold 8 weeks moist
|
Average
|
52%
|
From 25 different
experiments
|
The estimate rose from
42% to 52%. I then used this estimate for the Rule of 3s.
x=
|
0.52
|
|
n=
|
3
|
/
|
n
|
|
0.4
|
|
n=
|
3
|
/
|
0.52
|
|
|
|
n=
|
|
|
5.76923077
|
Figure 2. Spreadsheet “machine”
to calculate sample size using a percent estimate. In the blue box I input the
estimate percent. The orange box returns the suggested sample size. I obtained the percent estimate from table 3 above.
For 52% estimate
germination, and rounded up to a whole seed, I need 6 seeds per cup for the 24
cups. Although I didn’t have enough for 8 seeds per cup, I do for 6, but I will
use 7 to use all my seeds, and more is better.
In conclusion: I will plant 7 seeds per cup, 24 cups per each species,
on one shelf.
Future
I now must designate
the cup distribution. Cups should be distributed across the shelves randomly.
This will be for another discussion. Updated: added:
Figure 3. Randomized placement of Baptisia cups on shelf for 2017 planting.
Utilizing this concept I
will perform a similar method to obtain how many seeds for the carnivorous
plant portion of my experiment.
Thank you for viewing.
Comment if you want. Enjoy the day!
Citations – Power
Analysis
Citations – Baptisia germination literature
Boyle, Thomas, Kristen Hladun. 2005.
Influence of seed size, testa color, scarification, method, and immersion in
cool or hot water on germination of Baptisia australis (L.) R. Br. Seeds.
HortScience 40(6):1846-1849.
Bratcher, Carlma. 1990. Methods of
producing field-grown specialty cut flowers. Oklahoma State University,
Sillwater: 18-23.
Bratcher, Carlma, John Dole, Janet Cole.
1993. Stratification improves seed germination of five native wildflower
species. HortScience 28(9):899-901.
Dirr, Michael. 1987. I gave it 30
minutes, I couldn’t relocate the exact citation, too bad. Probably came from: The reference manual of woody plant
propagation: from seed to tissue culture, 2nd edition. But I
could be wrong. I will leave it in this post with this note. It did not change
the cold stratification results.
Estes, Becky, John Kush. 2007. Longleaf
pine: seeing the forest through the trees, proceedings of the Sixth Longleaf
Alliance Regional Conference; November 13-16, 2006, Tifton, GA. Longleaf
Alliance Report No. 10.
Ksiazek, Kelly, Jeremie Fant, Krissa
Skogen. 2014. Native forbs produce high quality seeds on Chicago green roofs. Journal
of Living Architecture. 1(2):21-33.
Voigt, John W. 1977. Seed Germination of
True Prairie Forbs. Journal of Range Management, 30(6):439-441.
Wennerberg, Sarah. 2005. Propagation and
field assessment of West Virginia native species for roadside revegetation.
Eberly College of Arts and Sciences West Virginia university Thesis.
