klaflamme's blog

Mechanical damage by cutting leaves has shown less response in plant defenses than herbivore damage because enzymes released by herbivores trigger a greater response. Mechanical damage increases concentration of jasmonic acid less than herbivore damage (McCloud and Baldwin 1997). The amount of trichomes increases slightly with clipping leaves, but not like damage from herbivores (Björkman et al. 2008). Applying jasmonic acid has been shown to increase trichrome growth without damage from herbivores. Trichomes are hair-like defensive structures that grow on the leaves of plants that impede movement. Jasmonic acid induces creation of secondary defensive compounds that are less favorable to herbivores and impede herbivore growth (Tian et al. 2012). Mechanical damage with treatment of jasmonic acid may produce the best results for herbivore deterrence.

The reaction of plants to stress from their environment involves a series of pathways which induce defenses (Tian et al. 2014). These pathways involve several hormones which trigger the defense responses in plants. Jasmonates are stress-induced phytohormones that incorporate biotic and abiotic cues that regulate plant growth, development, and defense responses. Jasmonic acid is a hormone plants release to control the responses from herbivore consumption. While impeding growth, the application of jasmonic acid to plants has been shown to enhance the treated plant’s natural defenses (Huang et al. 2017).

On February 21, 140 surviving plants were transplanted to one-gallon pots using Pro Mix BX general purpose soil with mycorrhizae (Premier Horticultural Inc., Quakertown, PA) and treatment fertilizer was added (Osmocote Classic 14-14-14, Scotts-Sierra Horticultural Products Company, Marysville, OH). Treatment and fertilizer amount was randomized, 70 plants were assigned treatment and 70 were not. Half the plants in each treatment group received one teaspoon of fertilizer, the other half received two. After application of fertilizer, plants were put onto greenhouse bench and watered. Plants were watered daily. 

Plant height was then recorded on surviving plants on March 7. Dead plants were discarded. Treatment of mechanical damage was done by cutting off half the new leaflets and one spray of 0.5mM solution of jasmonic acid in acetone was applied. Control received one spray of acetone and no mechanical damage. Plants were placed on a bench in the greenhouse for 14 days so treatment could take effect. 

On March 21, herbivore preference was tested. One new leaf from the treatment group and one from the control, both having received the same amount of fertilizer, were put into a 15 cm petri dish with one tobacco hornworm Manduca sextaordered from Great Lakes Hornworm. There were 84 replicates of herbivore preference experiment. A photo of the leaves pre-hornworm was taken, then again after 18 hours of hornworm exposure. LeafByte was used to determine percentage of each leaf consumed. On March 28, height, root and shoot biomass were measured, and flowers per plant were counted. 

Experiment began January 19, 2019. Tomato species Lycopersicon lycopersicum, a dwarf variety that matures in 50 days, was used. Seeds used were “Tumbler Hybrid Tomato” (Lake Valley Seed, Boulder, CO). Seeds were sown into two 128-plug flats with a moist seed starter soil (Organic Starter Premium Potting Mix, Epsoma, Millville, NJ). Flats were placed on a greenhouse bench with natural light. Germination was induced by placing onto a heat mat. On February 5, 100 ppm of 20-10-20 fertilizer was added at every watering. Fertilizer was increased to 200 ppm after 7 days of growth.

Sieve Cells are called this because the tubes act as one continuous cell. Sieve tubes are a collection of sieve tube elements that tend to look empty, have very prominent joints, and are very long. They don’t have a nucleus (like xylem, but are living cells), don’t have a cytoskeleton, have very few organelles  (no Golgi, plastids or lignin; some ER, few mitochondria, have plasma membrane). There are companion cells next to sieve cell-ordinary cell that feeds sieve cell/keeps it alive. A ‘Sieve plate’ is the wall in between each cell. Sieve plates are almost like slime/goo and are very prominent-these are phloem clots (like a blood clot in a plant) stops the flow of nutrients. This is how plants keep nutrients from being lost if it is damaged/cut. They contains special clotting proteins and polysaccharides.

Covariates are continuous and natural, a block is discrete. Blocks need one of each treatment and must be decided before experiment is started. Blocking affects how replicates are assigned to treatments. Thus, once a decision has been made, it cannot later be decided to remove ‘block’ as a term in the analysis (some people would today if it does not explain significant variation). The decision to measure covariates can be made at any point. Many potential covariates can be measured. If the analysis shows that they do not explain significant variation in your response, they can be dropped from the model.

 A split-plot design is defined as one factor assigned at the level of main plots (water or none), which are then subdivided into subplots that are the level of replication for a second factor (corn type). A nested design is defined as a research design in which levels of one factor are hierarchically subsumed under levels of another factor. As a result, assessing the complete combination of A and B levels is not possible. Like a split plot design, replication differs for different factors within the same experiment. However, in a nested design each level of one factor is NOT crossed with all levels of the other factor. An example we could use to understand this is to compare differences between 6 cities, 3 on the west and 3 on the east. We can’t analyze the interaction between city and coast because Boston is only on the east coast, Seattle is only on the west.

A Type I Error is a false positive where a true null hypothesis is rejected, and a Type II Error is a false negative where a false null hypothesis is accepted. In this case, a Type II Error would conclude this pesticide does not increase cancer rates when it does, a Type I Error would conclude this pesticide does increase cancer rates when it doesn’t. We would want to minimize our risk of Type II Error because this sort of error could mislead people into thinking the pesticide is safe when it is not, and the continued use of it would increase cancer rates. A Type I Error is less important because in this case, cancer rates would not rise since the pesticide would be deemed unsafe. It may cause farmers to lose millions of dollars, but that is less expensive than the lives of people and the compounded medical costs of cancer of the country.

The randomization procedure to choose the territories may give us more large territories than small or vice versa. It could also generate coordinates that are close to one territory, both decreasing the randomization. A better randomization procedure would be to assign a number to each possible territory, (1-10 for the large territories, 11-20 for the small territories, for example) that could be used and use a random number generator, then generate 5 numbers for larger territories and 5 for smaller ones. This way, there is an even amount of territory sizes and each territory experimented on is randomized. 

The factors that determine power are Replicants, Experimental noise (random variation) Experimental design (the various aspects, but a simpler design has more power), P-Value cutoff (usually P = 0.05), and How strong the treatment effect is (how big the difference is between treatment and control mean; stronger effect = more power). The power of our experiment is 84.2% (842 successful/1000 total). The probability that we get valid results is 0.842. The probability that we don’t get these results is 1 minus the power of design; 1-0.842=0.158, a 15.8% chance that water temperature doesn’t affect mating behavior. This would be a Type II error because a Type II error is the probability of falsely concluding there is no effect of treatments, when there really is a treatment effect. The probability that we would incorrectly conclude that water treatment did have an effect would be 0.05, since this is the cutoff for a significant alpha value. Since we would be rejecting a true null hypothesis, this would be a Type I error.