The coat color of mice can be black, white (albino) or agouti. This trait is controlled by two genes - one that controls "agouti" vs "black" coat color (we'll call this the "coat color gene (locus)") and one that controls "albino" vs "not albino" (we'll call this the "albinism gene (locus)"). "Agouti" and "not albino" are dominant. These genes interact (epistasis): if a mouse expresses the albino trait, it is albino regardless of the alleles at the other locus. This is because it does not produce the pigments necessary for either agouti or black coloration.
You receive a set of pregnant brown mice. You know that the mothers are doubly heterozygous (heterozygous at both the coat color and albinism loci) and that the fathers were albino. All of the fathers have the same genotypes (i.e., your results must be explainable by a single cross). When the mothers give birth, you find the following coat colors in their offspring:
145 agouti mice
153 black mice
302 albino mice
The traits we are studying are NOT sex-linked or associated with a lethal condition that would increase the death rate of particular genotypes before the offspring are born.
Your job is to determine the parental genotypes. To do this, you will use a Punnet square (or similar notation) to generate expected phenotypic outcomes for this cross under each of several possible parental genotypes.
a) Father was albino and heterozygous at the coat color locus.

b) Father was albino and homozygous black at the coat color locus.

c) Use two chi-square tests to evaluate the support for the null hypothesis that the observed distribution of offspring is the same as that expected for the paternal genotypes in a and b. Indicate for each hypothesized set of parental genotypes whether the deviation between the expected and observed outcomes exceeds the 0.05 "reject" level. Can any of these possibilities be eliminated from further consideration?

d) Why do we know that the fathers were not homozygous agouti?