I am presuming that by "k is first" you mean k is the smallest integer in T and thatBecause you are using Hammack's work you surely can follow this proof.
Suppose that \(\displaystyle P\in\mathbb{Z}^+\) and \(\displaystyle (\forall n\in\mathbb{Z}[P\ne n^2)\)(i.e. P is not a square).
You should know about the floor function: thus \(\displaystyle 0 < \sqrt P - \left\lfloor {\sqrt P } \right\rfloor < 1\) (**)
O.K. here goes: Suppose that \(\displaystyle \sqrt P\) is a rational number.
Then there are two relatively prime positive integers \(\displaystyle a~\&~b\) such that \(\displaystyle \sqrt{P}=\dfrac{a}{b}\) or \(\displaystyle a=b\sqrt{P}\)
Define a set \(\displaystyle T=\{n\in\mathbb{Z}^+: n\sqrt{P}\in\mathbb{Z}^+\}\). \(\displaystyle T\) is not empty because \(\displaystyle b\in T\)
So \(\displaystyle T\) is a non-empty subset of the positive integers as such \(\displaystyle T\) contains a first term. Call it \(\displaystyle k\).
From (**) we get \(\displaystyle 0<k\sqrt P-kH<k\). BUT \(\displaystyle (k\sqrt P-kH)\in\mathbb{Z}^+\). Why??
Moreover, \(\displaystyle (k\sqrt P-kH)<k\) and \(\displaystyle (k\sqrt P-kH)(\sqrt P)\in\mathbb{Z}^+\).
That violates the minimality of \(\displaystyle k\),. Its a contradiction to supposing \(\displaystyle \sqrt P\) is rational.
To restate: If \(\displaystyle P\) is a positive integer that is not a square then \(\displaystyle \sqrt P\) is irrational.
[MATH]H = \lfloor \sqrt{P} \rfloor \implies H \in \mathbb Z^+.[/MATH]
If those guesses are wrong, I do not get the proof.
If they are correct, it's a very nice proof.