Third week of the experiment

Thursday, July 18th, 2013

Third week of the experiment; about halfway through!

When we came back in the field this Monday, we noticed some drastic changes among the kale plants:

Most of them had almost doubled their leaf number and size and we also noticed a much reduced presence of flea beetles.  The new kale leaves were not only less attacked by herbivores, but also much thicker and more resistant to other pests.

In order to adapt our treatment as kale was growing, we decided to apply more silk per plant (about 30 strands per plant).

As a consequence, the silk strands were much more visible both on the frame and on the treated plants on the day of the application.  Thus, we became interested in seeing for how much time was the silk remaining on the treated plants.  Generally, it was very difficult to see the silk again on the treated plants one day after we applied it on them.  Also, after we finished the silk treatment for the 36 plants, a heavy rain started to fall and winds started to blow, probably reducing silk’s persistence on the leaves.

Herbivory

In order to have an idea of the herbivory level at the stage of the experiment, we decided to create a system of visual reference for us to evaluate the overall damage on each leaf of each plant.  We decided to work with this ‘visual approach’ because damage and leaf shape were generally not regular, making an approximation of the ‘missing parts’ difficult.  Also, having a reference table to estimate the level of herbivory for each leaf made our task easier and less time-consuming than more extensive calculations.

Using this table (part of it shown here), we proceeded to an estimation of an average of total damage for each plant from each leaf’s percentage damage value (1 corresponding to 0-10% damage, 2 to 10-20% damage, up to 9, corresponding to 80-90% damage).

This method contains both advantages and disadvantages:
Advantages
-easy to use
-time-efficient
-quick, time-efficient technique giving a general idea of the state of the experiment; very helpful for troubleshooting
Disadvantages
-subjective to the person using the table
-less accurate than direct measurements
-difficult to perceive the difference between close categories (e.g.[50%-60%[ and [60%-70%[)
-most new leaves show low levels of herbivory and older ones show high levels of herbivory: those extreme results 'cancel' themselves out in the calculation of the average damage for each plant

Preliminary results using the reference table

First of all, it is important to mention that different sizes of samples are available to compare the results between the control and treated plants. Therefore, comparing groups of different sizes creates a bias in our experiment.  
The following graphs show a very general overview of the results for the level of herbivory at this stage of the experiment. From this, we can see that most of the data is very concentrated below the [30%-40%[ interval. While being visually easy to understand, those 'box and whisker' diagrams do not show clearly the differences between herbivory on spider silk treated and control plants.

The results of our preliminary statistical analysis do not give us confidence to conclude regarding the level of herbivory for treated and non-treated plants.  A similar analysis shall be made at the end of the experiment along with an analysis of variance to see if we can or not perceive difference regarding the level of herbivory among both treatment groups.