Good question. As far as i know you need to run more trials to determine if anything significant is going on. As an example, let's say you have seconds 1-10. And you have trial 1 and trial 2. trial 1 is control, trial 2 is treatment. You then measure water content every second (from 1 to 10). Now you have 2 columns (control and treatment). But for analysis, you can only compare the datum at second 3 in column 1 to datum at second 3 in column 2. And you can't run stats on that (2 data points). So you would need to repeat this process several more times (depending on the effect size). If you did this 6-10 times you would then have 6 data points at second 3 for control and 6 at second 3 for treatment. You could then do a simple t test between each second to determine if there were any differences at each second. If you added a second treatment rgoup (column 3) this then has to be an ANOVA, not a t test.

Do this make sense? As it is you have some pilot work that can help point you toward what the study needs to be. (In other words, plotting difference at each second vs. time might give a nice graph suggesting what might happen with more trials, but it doens't tell you anything statistically).

Let me know if this makes sense.