Statistics question !!!

Suppose you are interested in following two independent traits in snap peas – seed texture (S = smooth, s = wrinkled) and seed colour (Y = yellow, y = green) – in a second-generation cross of heterozygous parents. Mendelian theory states that the number of peas classified as smooth and yellow, wrinkled and yellow, smooth and green, wrinkled and green should be in the ratio 9:3:3:1. Suppose that 100 randomly selected snap peas have 56, 19, 17, and 8 in these respective categories. Do these data indicate that the 9:3:3:1 model is correct? Test using a = .01.

Could someone help me with the null and alternative hypothesis? How do you go about answering this?

We'll use a Chi-Square Goodness fo fit Test.

This test uses expected values of the combination of variables and compares them to actual values.

First we generate expected values.

smooth yellow 56.25
wrinkled yellow 18.75
smooth green 18.75
wrinkled and green 6.25

The null hypothesis is that there is no significant difference between observed and expected value--which means the two traits are independent in their appearance so therefore they are on different chromosomes.
.
Chi^2 = (56.25-56)^2/56.25 + (19-18.75)^2/18.75 + (19-18.75)^2/18.75 + (6.25-8)^2/6.25 =

0.0011111110 +.003333333 +0.163333333+ 0.49=0.657777778


The test has number of traits 4 -1 = 3 degrees of freedom

1X2 with 1 degree of freedom is 11.345.

Our Chi-Square test statistic is 0.65777 < 11.345. So we do not reject null hypothesis.