Reptile & Amphibian Forums

I am a senior at Jefferson County International Baccalaureate School and I am doing a science fair experiment in which I am comparing the theoretical outcome of corn snake clutches to the actual outcome by looking at the breeding of a homozygous corn to a heterozygous corn (for any trait/color/pattern. For instance, say I bred an amel to a het amel, a punnett square for this that says 50% of the offspring should theoretically be amel. However, in an actual clutch 3 out of 10 offspring were amel. I was wondering if you anyone could give me clutch information to help out. I would just need to know the overall clutch size and then how many were phentypically amel./lavendar/anery/etc.. and how many were not. Thank you very much!!!!

I don't mean to be rude or offensive but I was wondering something, If you got the clutch information about the snakes (corns) being a certain color (phenotypically & genotypically) then how would it be your experiment? Wouldn't it be you just explaining what someone else did after maybe close to a decade of work? Just wondering.

no. i am analyzing the statistics of corn snake breeding, not the actualy breeding. in other words my project is based on a posteriori data. although it is technically a science project it is more mathetmatically based and will entered into the math division of the science fair contest. it is perfectly exceptable and normal to take already conceived data and analyze it for the type of project i am doing. however, in order to get data i have to have willing donors

There are a few other things you may consider before assuming that your punnet square predictions are faltered or that the genotypes of your snakes were incorrect.

1st it is possible that there was an uneven, spontaneous die-off of embryos in-vitro. Suppose you originally had 7 amel and 7 other embryos post egg-laying, but 4 spontaneously aborted (and were reabsorbed). If by chance they were all amel you would be left with a clutch of 3 amel and 7 other in a clutch of 10, instead of 7 amel and 7 other in a clutch of 14 as predicted by punnet.

2nd, do not discount the possibility of chromosomal crossovers, they are a common occurence, built-in to add genetic diversity to a population.

3rd, statistics, statistics, statistics. This is but one clutch of eggs. The theory invovled in punnet squares was developed from very large sample sizes. If you were to have bred 4 pairs of snakes with 10 eggs per clutch, perhaps the percentage of amels would be 30% for one, 70% for another, 40% for the next and 60% for the last, averaging out to 50% for a large population. Obviously, this is a simplified case and more like 20-100 clutches would be needed to achieve statistical significance to verify a punnet prediction. It is apparent why the first genetic studies were done with pea plants (e.g. cheap, fast generation time, low maintainence).

I hope this helps you piece together your breeding conundrum, and appologize if at any times it sounded preachy

i had not thought about the being "reabsorbed" issue, but i will be sure to include it in my paper now. also, genetic crossing over is discussed in my paper and i am not focusing only on one clutch. i already have 38 clutches to work with (over 300 individual snakes) and i am using the averages as my basis. i am actually testing mendel's proposal/observation that the more data an individual has the closer the individual will come to the theoretical value for the data so right on, buddy!

Ah I misunderstood. Gotta love stats.

The reabsorbed issue only pertains to embryos that are terminated early enough to be worth the enregy of absorbing (e.g. a larger or already shelled egg might be laid as a dud along with the rest instead of broken down by cellular and chemical means at a greater cost).

It sounds like you are compiling a pretty detailed report so I will add that a smaller (and therefore less likely to upset genotype and phenotype) concern is that of point mutations (usually cause no noticeable difference although can be severe) and chromosomal abnormalities (e.g. forms of albinism caused by improper encoding of enzymes used in the formation of melanin).

Either way, 38 clutches sounds like more than enough.

Can you perform a regression to come up with an equation that predicts how close an observed population of offsrping will come to theory based on the size of the population?

Just some curisoity from a fellow science dork (currently embroiled in torturous gross anatomy and histology classes ).

right now i am planning on use the chi square method of hypothesis testing in conjunction with standard deviation and histogram formulas. is there another formula that would work as well? let me know if you know of one

So, say you made a graph. On the y axis you have the average percentage of a clutch that is amel. On the x axis you have the number of clutches used to compute the average. As more clutches (or snakes) are added the average percentage amel should approach 50%.

Likely, this relationship is linear, and therefore you could write an equation y = mx b, where you could predict how well a given accumulation of clutches would agree with theory (by closeness of y to 50%) using only the number of clutches (or snakes) in the population and the constant slope m (b would be 0 in this case).

With the equation you could produce a graph for all situations out to an infinite number of clutches (or snakes). I think it might be a nice visual to summarize your project's goals or observations.

Test it out with your data, e.g. plot the points first considering 1 clutch (or snakes), then 2 clutches, then 3 and so on (excell will make this much easier) and see if the data seems to be linear.

Good luck, and if you need any help working this out in excell, don't hesitate to ask.

exactly how do i get excel to do this for me?
by the way, i will be leaving for vacation in the morning and won't be back until saturday, however, you can continue to reply on here or you can email me-- greentangerine07@hotmail.com thanks so much for your help!

Ok so here's the first thing to do.

calculate the avg percentage amel value for

Clutch A
Clutch A B
Clutch A B C

an so on to however many clutches you have, these will be the y values.

Your x values will be the number of clutches used to calculate each y value (e.g there will be an x value paired to every y).

Next, open a new spreadsheet.

In column A place your x values, one in each cell.

In column B place each y adjacent to it's corresponding x value.

E.g. if cell A1 is 1 clutch (1 clutch was used to calculate y) then cell B1 should contain that y value.

Once you have the two columns filled with your data, drag to select all the data in the two columns. They will highlight.

In the command bar at the top there should be an icon with what looks like a bar graph on it, it will say chart wizard when you highlight it, click on it. If the icon is not there, go to view, toolbars and select standard and formatting.

When it opens you will be asked for a chart type. Under the standard chart types tab, choose XY (scatter). click next.

You should see the points plotted out. Hit next again give it a title and name the axes. Hit finish.

Your graph will not be displayed among the cells.

Click on a point. All should be highlighted. Right click on a point and select add trendline.

There are many types to choose from. Linear might work.

In options highlight display equation on chart and display R-squared value on chart (I believe R-squared is from the chi square method you talked about earlier, if not then I'll tell you that the closer R-squared is to 1 the better your data fit the imposed line generally higher than .95 is comsidered a clean fit).

I've been thinking about the nature of your data and it strikes me that 50% amel by these calculations will never be achieved. I think your data will appraoch an asymptote at 50% amel. For this reason the linear equation might not be the best way to represent your data. In the trendline menu you might try fitting a logarithmic line to the data and see how the R-squared value looks.

Anywho, good luck and have a restful vacation.

hey! just got home. i still have a little more data to collect before i can get numbers/averages calculated. however, when i do would like me to email you a copy of my paper...you could proof it/comment/give me feed back if you like. anyways, thanks for ALL your help! -paige

Be glad to give it a quick look over. Used to be a writing tutor in college. I don't want to give my email out too publicly so I made a temporary one.

ECP-Qroberts@hotmail.com

Good luck.

Hi

Some info from my records for (Pure corn to pure corn) mateings for 2004.
unfortunatly i dont write the dates down for my records, i have egg boxes ready for the eggs with the dates writen on lables which are thrown away when i clean the boxes after the hatching process.
laying and hatching times have been quite normal for nearly all my snakes.

All eggs These eggs were incubated at a constant 85 with a humidity content of 96%.

male albino (a) x female het/albino (a):

1st clutch = 17 eggs with 1 bad = 16.
all 16 hatched - 13 amel and 3 het - male to female ratio = 7m 9f.

2nd clutch = 13 eggs all good.
all hatched - 9 amel and 4 het - male to female ratio = 5m 8f

male albino (a) x female het/albino (b):

one clutch = 21 eggs all good.
all 21 hatched - 17 amel and 4 het - male to female ratio = 7m 14f

male het/albino (a) x female albino (a):
1st clutch = 10 eggs all good.
all 10 hatched - 6 amel and 4 het - male to female ratio = 7m 3f

2nd clutch = 6 eggs all good.
all 6 hatched - 5 amel and 1 het - male to female ratio = 4m 2f

male het/albino (b) x female albino (b):
one clutch = 19 eggs all good.
18 hatched and 1 died in the egg = 18 - 10 amel and 8 het - male to female ratio = 9m 11f. just for the record the dead one was a het.

and one Het to Het mateing with results of 15 good eggs. all 15 hatched - 5 amel and 10 het - male to female ratio = 3m 12f.

All other mateings with Pure corns i done in 2004 were 100% x 100% trait to trait so all results were 100% trait possitive with the exception of some crazy patterns that popped up.

I hope this small veiw of my breeding results help you out with some of the info you need. and im sure there are a few more people on this forum that can help out with some results...

Paul..
-----
Pinky's taste nice,only if there fresh..

The important lesson you're learning is that nature doesn't conform to abstract concepts like statistical likelihood.

A 3:10 ratio isn't noteworthy in the pairing you describe. You could also breed a homo Amel with a het Amel and get all Amels - nature doesn't read the manual.

Clutch predictions are just that - they're informed by years of exprience, but they're still just a best guess.

Take a coin. You prove beyond doubt that it is not weighted or fixed in any way. You toss it 100 times - it comes up heads each time. Each time you toss it, the statistical likelihood is 50:50 that it will come up tails. No matter how many times in a row it comes up heads, it still has a 50:50 statistical likelihood of coming up tails next time.

that's exactly the point im trying to observe/test/evaluate through my project

To be honest, it's already been observed, tested, proven and accepted. Your research will be your own, but what you'll achieve is the proof of existing knowledge.

Sounds like a fun project anyway. I wish I could have tied my studies into a subject that I actually had a passion for!