To know a plant, grow a plant!
Refresher on the Experiment: Standard Fast Plants were planted in quads (128 plants total) and data was taken approximately every 2 days on the plants' heights. Prior to the day of flowering, the top 15% of the tallest plants were chosen to be inter-mated, and the other 85% of plants were discarded. The top 15% of tall plants were mated and allowed to set seed. The Offspring generation was planted from those seeds. Data was taken on the Offspring plants' heights in the same manner as for the Parent generation. Height data for the two generations was compared to determine if an increase of plant height had occurred from our selection for that characteristic.
Results, briefly: We did not see an increase in average plant height between our Parent and Offspring generations. See below for Height Histograms for Parent and Offspring generations.
Discussion of Results: There are many factors involved in the expression of a plant's height characteristic, including temperature, light intensity, depth of the planted seed within soil, and water availability. Each of these factors will be discussed below.
Light Intensity - Though we did our best to control for similar levels of light intensity by raising or lowering our quad-shoeboxes to keep plant apices within 5cm of light, complications arise in the fact that if one plant in a shoebox (8 quads, and 32 plants per shoebox) is significantly taller than the others, only that plant will truly be within 5cm of light... the other plants in the shoebox will be receiving lower levels of light which may cause elongation and skewing of height data.
Temperature - Our location for growing plants is under a light rack in the basement of our building, and our basement is highly susceptible to temperature fluctuations throughout the seasons. Our Parent generation was grown during the winter/early spring season during which our basement ranges from quite chilly to almost uncomfortably warm (we did not take temperature data). Our Offspring generation was grown during the summer during which our basement ranges from approximately room temperature to quite hot. It is likely that each generation may have responded differently to its differing temperature conditions... another potential factor for skewing of height data.
Depth of the planted seed within soil - Though we try our best to be consistent in planting (we all discuss our planting protocol before hand), there is likely still variation in the planting style of individual team members. If one person was to plant their seeds deeper in the soil than another person, the shallower seed may have emerged from soil first, giving it an advantage in height over other plants (which must be considered for plants with as fast of a life cycle as these!).
Water Availability - Our quads were watered automatically on a wicking system from the shoebox reservoirs and supplemented with top-watering to maintain wick-water-continuity. However, variance between team members watering protocols (we did not standardize volumes of top-watering) or variance in position of wicks may have led some quads to receive less water than others. Some plants are very sensitive to water stress and will modify their allocation of root/shoot growth when they don't receive enough water. We have no way of knowing if all quads were equally hydrated for the lives of the plants and this may have skewed our height data.
On another note: Despite the fact that our experiment did not show an increase in plant height between generations, it is established that selection CAN be successful upon plant height. For example, in working with NASA to put Fast Plants in space, Paul Williams, "The Father of Fast Plants" was able to make significant gains in reducing the height/stature of Dwarf Fast Plants while producing the Astroplants stock. Perhaps we would have been more successful if we had selected for the 15% shortest plants instead of the tallest.
Moving Forward: This Artificial Selection project is being repeated (see Artificial Selection II) but with selection for a plant's salt tolerance rather than height. It is expected that a selection for salt tolerance will be less susceptible to environmental factors like temperature, light intensity, and depth of planting of seed. Efforts are currently underway to make our results for Artificial Selection II available on this site. Keep checking back to see what happened!
Did you verify your data also with ANOVA statistical test?
We did verify our data by comparing average heights in the parent and offspring generations using an ANOVA test. The tests returned interesting results due to our methods of processing the data. The explanation is a bit lengthy, but I'll try to be clear... feel free to inquire further if anything is unclear or confusing.
In order to construct the histograms above, we had to create a single value for plant height, but we wanted to calculate this value from multiple days of height data per plant in order to create height values that considered a plant's growth across multiple days instead of at a single point in its life.
Unfortunately, while collecting height data on our parent and offspring generations, we were unable to collect data on consistent days. Our parent generation heights were calculated from the average of height data taken 5 Days After Planting (DAP), 7DAP, and 10DAP. MEanwhile, our offspring generation heights were calculated from the average of height data taken 4DAP, 7DAP, and 11DAP. The histograms above were calculated using those methods.
A problem arises in the fact that by 11DAP, a number of our offspring generation had begun rapidly increasing in height (bolting) as they entered the flowering stage of development. To account for this, we calculated individual plant heights using a new methodology by extrapolating height data. For the parent generation, we extrapolated 4DAP heights from data taken 3DAP and 5DAP. For the offspring generation, we extrapolated 10DAP heights from data taken 7DAP and 11DAP. Thus, we were able to calculate a second set of histograms for each generation using data from 4DAP, 7DAP, and 10DAP for both generations. It should be noted that extrapolating backwards from 11DAP to 10DAP only lessens (but does not completely negate) the discrepancies introduced by rapid height increase from bolting.
The two sets of histograms (one from normal data and one from extrapolated data) reported similar results: no notable increase in average height distributions between generations.
The interesting part of the data comes from ANOVA tests of plant heights. When conducting an ANOVA test using normal plant data (non extrapolated) we returned a p-value of 0.054, very near to what we would have considered significant for an increase of height between generations. However, when conducting an ANOVA test using extrapolated data, we returned a p-value of 0.691, much less significant.
With such huge discrepancy between the two calculated p-values, interpretation of the data is difficult using p-values alone. We opted for the histogram analysis of our data, where it is more clear that our average height distributions did not significantly increase between our parent and offspring generations.
Any comments you have on this situation or on our interpretation of data would be very interesting to us.
I think you did the best with your data set... I will share with you our results as soon as possible.
Our experimental design is quite simpler (I'm working with very young students and we will also chose a graphical approach to analyze the results)...We decided to select the highest and the smallest and only after witch kind of seeds to replant... (sorry for my "bad" english)
Thank you for sharing!! Great results, looks like your selections resulted in a significant reduction in height, congratulations!