1
$\begingroup$

I wish to compute class group of $K = \mathbb{Q}(\sqrt{30})$. I have checked, using sage, that the group is $C_2$, but I have the incorrect answer of $C_2\times C_2$. I outline my line of reasoning below; I would appreciate some guidance on how to correct my approach. To save some reading, I seem to be making a mistake when it comes to checking if $P_2P_5$ is principal or not.

The Minkowski bound is $\sqrt30$, and so I've checked how the rational primes $2$,$3$ and $5$ factorise in $K$, with $\alpha = \sqrt30$

$$(2) = (2,\alpha)^2 = P_2^2$$ $$(3) = (3,\alpha)^2 = P_3^2$$ $$(5) = (5,\alpha)^2 = P_5^2$$

I have checked that these prime ideals are not principal. We see that the orders of all the elements in the class group are $2$, and $P_2P_3=(6+\alpha)$, so $P_2$ and $P_3$ are in the same ideal class. Since I should be getting a class group of order 2, I expect something similar to happen when computing $P_2P_5$, yet if this were to be principal, generated by $x+y\alpha$, say, then, since $N(P_2P_5) = 10$, we need integer solutions to: $$x^2 - 30y^2 = \pm10$$ A number theoretic argument, or a check on Wolfram Alpha, shows that no such solutions can be found. This has led me to believe that $P_2$ and $P_5$ are not in the same ideal class (else, since the order of the ideal class is $2$, the product of the ideals would be principal), and hence that the class group is $C_2\times C_2$.

$\endgroup$
13
  • $\begingroup$ Looks like a copy from the question in the physics forum. $\endgroup$ Commented May 26 at 13:43
  • 1
    $\begingroup$ right, the positive values (primitively represented) along the "river" are just $1,6.$ Climbing one level, new value $19.$ That's it, no $10$ $\endgroup$
    – Will Jagy
    Commented May 26 at 14:16
  • 1
    $\begingroup$ and $5x^2 + 10 xy - y^2$ starts $5, 14, 21,26,$ so there was no $-10$ either $\endgroup$
    – Will Jagy
    Commented May 26 at 14:22
  • 1
    $\begingroup$ WE, yes, for real quadratic it is two to one. $\endgroup$
    – Will Jagy
    Commented May 27 at 18:31
  • 1
    $\begingroup$ yes, in that case one to one. Here we need to take form discriminant fundamental; if we multiply some discriminant $D$ by some square, the class number usually increases, while the field has not changed. This is all in Cohen, A Course in Computational Number Theory. Chapter 5 is Algorithms for Quadratic Fields, begins page 223. Fundamental Discriminant defined page 224; then section 5.2 is Ideals and Quadratic Forms. I learned this material in D. A. Buell, Binary Quadratic Forms $\endgroup$
    – Will Jagy
    Commented May 28 at 17:05

1 Answer 1

1
$\begingroup$

*Note to self (and maybe others): $N((\beta)) = |N_{K|\mathbb{Q}}(\beta)|$, with a large emphasis on $||$

$P_5 = (5-\alpha)$, and so is principal.

$\endgroup$
1
  • 3
    $\begingroup$ put another way, form $6x^2 - 5 y^2$ is integrally equivalent to $x^2 - 30 y^2.$ One can see how $6x^2 - 5 y^2$ represents the coefficient $1$ as $6-5$ $\endgroup$
    – Will Jagy
    Commented May 26 at 14:54

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .