Markers: Give 1 point for any reasonably correct solution (note some questions are is not assessed). Give the full point for any solution to number 3 which is even vaguely reasonable. 1. If Z+Z\alpha is a subring then \alpha^2 is certainly in the set. Conversely if \alpha^2 is in Z+Z\alpha then any power of \alpha is so is any linear combination of powers of \alpha. But this set is a ring (closed under +, times and additive inverses). In each of the sample numbers just calculate the square and see if it is in the Z span of 1 and the number. For instance \alpha = (1/2) + (\sqrt 5)i/2 gives \alpha^2 = (1/4)(-4 +2i \sqrt 5 ) = -(1/2) + \alpha If this were a lin comb of 1 and alpha so would 1/2 be and we would have integers m and n so 1/2 = m + n(1/2 + i\sqrt 5/2) = (n/2 + m) + n i \sqrt 5) so n=0 and m is a whole number contradiction. [ 1 point for mostly correct] 2. The darkened points include 2+ root(3)i and -3 + 2 root(3)i. These are perpendicular and all the others are an integer times 2 + root(3)i plus an integer times -3 + 2 root(3) i. Every element of the form m + n root(3)i can be obtained by adding one such element to one of the following elements: 0, a+root(3)i , b+root(3)i where a is -1,0, or 1 and b is -2, =1, or 0. So in every element of the ring is congruent to exactly one of these seven elements. Other similar solutions are possible. 3. The complex plane is 'tiled' by rectangles of area say A whose vertices are the elements of O. The ideal generated by a+bi is generated by rectangles of area (a^2+b^2)A. Roughly, there are a^2+b^2 little rectangles filling up each big one. 4. The condition on the function e: a+bi |-> a^2+b^2 is that for every pair of elements p and q of the ring O with q nonzero there are s and r so p = qs+r where r=0 or e(r) < e(q) This is the same as saying for any pair p,q the absolute value of p/q - s is less than one for some s in O. Any complex number is arbitrarily close to one of the form p/q (even taking q to be an ordinary integer) so this is the same as saying any complex number is at most distance one from an element of O. [1 mark for mostly correct] 5. The solution is basically given in the question. If this ideal I were principal then the number of cosets [O:I] would be m^2=3n^2 where m+ni is a generator. But 2 is not of this form so I is not principal. [ markers be careful to check studens say something at least slightly different from statement of question to get one mark] 6. Again the solution is in the hints. [Markers check they have added at least some comments of their own for one point] 7. Let's ask, what are the invertible elements in C(R,R)? These are functions f(t) such that there exists a g(t) with f(t)g(t)=1. These are the nowhere vanishing functions (which take either strictly positive or strictly negative values). If f(t) is a divisor of the sin function then f(t)g(t)=sin(t) for all t, where g(t) is a continuous function. It follows that the zeroes of f(t) are all integer multiples of pi. Now, if f(t) takes more than one zero, consider two adjacent zeroes a,b. We then have an interval [a,b] where a and b are integer multiples of pi, such that f(a)=f(b)=0 but f(t) is not zero if a < t < b. Choose any point c between a and b. I have f(c) is not zero and consider two functions defined g(t)= f(c) if t less than or equal to c g(t) = f(t)/f(c) if t greater than c h(t) = f(t)/f(c) if t less than or equal to c h(t) = f(c) if t greater than c. We have h(t)g(t) = f(t) obviously and h(t) and g(t) aren't invertible so f(t) is not irreducible. This proves irreducible divisors of sin take at most one zero and so the sin function can't be a product of finitely many. Of course then it does not have a unique factorization as a product of irreducibles, it has no such factorization. [markers: be generous! this is hard! 1 point for anything reasonable but note in class I made the statement that t divides into sin(t)] 8. This challenge question is not assessed. Let r be a nonzero element of R. Let's show it is contained in just finitely many ideals. Suppose r is contained in an ideal I. Then because ideals are principal we have an element s with I=Rs. Then r is an element of Rs so r is a multiple of s. We know R has unique factorization by a theorem in the course. So if the prime factorization of r is p_1p_2...p_n with the p_i possibly including repetitions, then all divisors of r are obtained by multiplying one of the 2^n subsequences of (p_1,...,p_n) and multiplying by an invertible element. The invertible element does not affect the ideal so there are at most 2^n ideals containing r. Note there could be fewer because some p_i may be an invertible element times another. The next part of 8 is a challenge question and I say "see me if you solve it." The question is, what extra properties does a principal ideal domain have to have in order to be Euclidean.