I have been trying to characterize integers representable by several ternary forms and reached a roadblock with this particular form: $$x^2+2y^2+7z^2$$
Ideally, I am looking for a characterization of square-classes represented. Computer calculations were unable to find any class number 1 ternary sublattices or an entire ternary genus. Any help with this problem or direction towards useful sources would be very much appreciated.
Very surprising to see a post on this, where someone knows what the words mean. Kaplansky wrote an early article on $x^2 + y^2 + 7 z^2$ in which he found it convenient to include your form. I'm afraid my websites are currently down with the host computer; anyway Math Comp 1995, The First Nontrivial Genus of Positive Definite Ternary Forms.
The entire genus is $$ x^2 + 2 y^2 + 7 z^2, \; \; x^2 + y^2 + 14 z^2. $$
Kaplansky called your form $k.$ He points out that Gordon Pall proved that $k$ represents all eligible numbers that are equivalent to $0,1 \bmod 3. $
alright, up to 1,000,000, and ignoring multiples of 49 (either form represents a number $n$ if and only if it represents $49n$) the comparison of the forms of the genus on eligible numbers comes out to:
Meanwhile, the numbers that are NOT represented by the full genus are $49^n (49m+21,35,42):$