Science Buddies: "Ask an Expert"

My science fair is determining if different meat products contain antibiotics by analyzing bacterial growth over time. I have put organic and non organic meat juices onto two different petri dishes with bacteria. However, because I am not using an incubator to grow my bacteria in, the growth rate is very slow. I was not consistent with gathering my data and counting the bacteria was very hard. How can I approximate my data if I counted 228 colonies on my final day of the experiment?

What an interesting science fair experiment. I'd like to help you with this, but I'm a little confused about exactly what you did. My understanding is that on the first day of your experiment you took meat juices from organic and non-organic meat and plated them on two different petri dishes that ALREADY had bacteria present? Then you counted the number of bacteria on these plates at different times. Even if you weren't very consistent about counting them regularly, you should know the initial bacterial count and the final bacterial count, so you can determine the difference at the end of the experiment to the beginning. If you only have these two timepoints, then that is what I would do. Compare the differences in # of bacteria between your two groups. Did one get more bacteria while the other got less? Did they both go up but one went up much more?

If you do have multiple timepoints, then you could compare each point on the graph as a percent of your starting value. For example, if your organic meat juice plate had 100 colonies to begin with, and the next time you counted it, it had 50 colonies, that would be 50% of the starting value. You would do the same thing for your non-organic meat (using the starting colonies for THAT plate). Then you could do a line graph for these presents to see them visually. Does that make sense?

I hope this helps, and if I'm not understanding exactly what you did in your tests, feel free to post again with more details, and we'll try to help.
JMP

Thank you for the advice. On the first day of my experiment, I took meat juices and plated them on different petri dishes WITHOUT bacteria in it. I checked the petri dishes for bacterial growth after 1-2 days but there was no growth so I had to wait until the 5th day of the experiment to actually see bacterial growth. (Given that I was not using an incubation system). My teacher gave me the suggestion of finding the surface area of the bacteria by finding the surface area of the petri dishes and counting the number of squares the actual bacteria take up. However, I'm not exactly sure how to do that because my experiment is done...

Ok, so what you did is plate the juices and count the number of bacteria that grew over time. When you say that the bacteria were difficult to count, is that because there were too many, or they weren't in individual colonies? The easiest thing to do would be to just graph the growth of the bacteria over time in a line graph. So on day 0 you would have 0 colonies, on day 5 you would have say 10, and by the end you would have 228. You could do this for both plates and then compare the growth.

If it is very difficult to count the colonies, then you could do two things. If it is just that there are way too many colonies to count, you can divide the plate into equal quadrants and just count two of them. Get the average, multiply by 4, and you have a fairly good estimate of the number of colonies on the plate. Alternatively, as your teacher suggested, you can try to look at the number of squares that have bacteria using a grid overlaid on the plate. However, as you said, the experiment is done, so you'll only be able to do that for the final day. That's okay though. Since both of your plates started with no bacteria, you know that that would mean 0 squares had bacteria. You wouldn't be able to do the in between timepoints, but you'd still have the beginning and the end.

Does that help?
JMP

Yes thank you. One more question. How would I do a standard deviation for my experiment? Do you think it would be possible? I have three trials: the first trial is testing beef juice with 10 Petri dishes. For every PAIR OF PETRI DISHES, I am testing non organic and organic. So for each pair, I have different variables I'm changing: For example the first pair will have change in pH, the second pair will be put in a cold temp, the third pair will be put inroom temp, the fourth pair will be put in the dark, and the fifth pair will have two swipes of bacteria. I am doing the same for trials 2 & 3 but instead using TURKEY juice for trial #2 and CHICKEN juice for trial #3. My teacher says we would need a maximum of 6 SAME data points in order to do a standard deviation. I'm not exactly sure if doing a standard deviation is possible with my experiment though.

Standard deviation is a measure variation between samples. I don't agree with your teacher that you need six samples to perform standard deviation calculations, but you definitely do need multiple replicates. Based on what you described in your last post, it sounds like you only have one plate for each condition (i.e. one plate that got non-organic beef juice in the dark and one plate that got organic beef juice in the dark). For an experiment, ideally it is best to try to do a minimum of 3 replicates for each condition, so you can eliminate the possibility that your results are due to random chance. So, in your experiment, you could pick one or two of your conditions and do multiples, for example a minimum of 3 plates getting non-organic beef juice in the dark compared with 3 plates getting organic beef juice in the dark. You would then compare the averages of the non-organic beef juice plate with the organic beef juice plate, and you could calculate the standard deviation of each group. As I understand the experiment you have described, you only have one plate for each condition, so you can't calculate a variation.

JMP