In pursuing a question of personal interest, I need to characterize solutions to the equation $m^2+9k^2\pm k=c^2$; in other words, I want to know when numbers of the form $m^2+9k^2\pm k$ are perfect squares. In the context of my problem, $c,k,m \in \mathbb N; 2<k<m^2$. By hand, I can grind out a few examples, such as $(m,k,c)=(4,3,10),(6,5,16),(6,7,22)$, and further convince myself that some values of $m$ (such as $1,2,3,5$) yield no solutions. But I want to understand as much as possible the nature of solutions, particularly with respect to suitable values of $m$. Apparently, not all values of $m$ yield solutions. I am not a programmer, so I have been limited to using a calculator.
I have tried simple restatements of the equation to see if I can perceive any avenues of attack, e.g.: $m^2+k(9k\pm 1)=c^2$ and $k(9k\pm 1)=(c+m)(c-m)$, but I am not having any insights.
My first question is: In the context of my equation, are there any general constraints on $m$? (Such as, but not limited to, constraints on its class with respect to cerain moduli. I can't see any such thing, but maybe somebody else can.)
My second question is: Is there any number, $m_0$, such that for $m>m_0$, the equation will always have at least one $(m,k,c)$ solution? Even if no particular $m_0$ can be identified, is it possible to decide if such a number exists?
I am most grateful for the thoughts of the community.
Above equation shown below has parametric solution:
$m^2+9k^2+k=c^2$
$m=[(5k+1)/4]$
$c=[(13k+1)/4]$
For, $k=(7)$ we get, $(m,c)=(9,23)$
For, $k=(11)$ we get, $(m,c)=(14,36)$
For, $k=(23)$ we get, $(m,c)=(29,75)$