Trend line in a graph illustrating photosynthesis rate

[zuzannamagdalena]

I am writing a report from an investigation whose aim was to determine the influence of CO2 concentration on the rate of photosynthesis in Elodea canadensis by measuring the amount of oxygen released by the plant at different CO2 concentrations.

I am required to include a trend line in the graph.

We were instructed the trend line should not go beyond the points of measurement. That is, if I am making measurements for 0.2%, 0.4% and 0.6% and include those values on the x-axis, my trend line should not extend onto 0.0% or 0.08%. However, when the teacher explained this to us, she referred to a situation where the trend line goes beyond the measurement points parallel to the x-axis. She said nothing about the y-axis.

I made a trend line using Excel. It looks like this (the graph on the left is the same graph without the trend line):
http://img524.imageshack.us/img524/1770/wykresy2.png

In the case of this graph, it looks a bit absurd because the number of bubbles cannot go below 0. However, it looks slightly less bizzare if you compare it to another graph I plotted for a different experiment:

http://img13.imageshack.us/img13/4610/wykresy.png

Here, the trend line goes beyond the measurement points parallel to the y-axis, too, but it looks more promising because the resulting y-values are realistic.

I am now wondering if the "do-not-go-beyond-the-measurement-points" principle applies to the y-axis, too.

If so, I don't change the settings in Excel to make a trend line that wouldn't go beyond the points of measurement, so I would appreciate it if someone explained how I can do it.

[JackBean]

Greetings to Poland

Of course, it applies only to x-axis. The point is, you have some values from x1 to x2 and you don't know, what will happen, if it would be smaller than x1 or higher than x2, there could be some limitation severely affecting the value. However, in the range x1-x2 you have measured it and you can say it's like this. Yet, you have some errors in your measurements, so of course the trend line will be up or down of it.

However, are you sure it should be linear? If so, I would exclude the last point. It's obviously too high (look on the error).

[zuzannamagdalena]

Thanks. Following your advice and my intuition, I handed in the lab report having left the graph looking the way it does. My teacher said it is OK.