The paper studies special mathematical objects called eta quotients, which are functions built from a classical function in number theory known as the Dedekind eta function. These eta quotients live in the world of modular forms, which are highly symmetric complex functions that play a central role in modern number theory (they were famously used in the proof of Fermat's Last Theorem). The specific eta quotients studied here are "irreducible cuspidal" ones, meaning they cannot be broken into simpler pieces and vanish at special boundary points called cusps. The main question is: how much information do you need to completely pin down one of these objects?
The first key result is that you do not need very much information at all. Specifically, an irreducible cuspidal eta quotient is completely determined by how quickly it vanishes at a surprisingly small number of cusps, roughly proportional to the logarithm of the level (a number measuring the complexity of the eta quotient). For a particularly nice class of these functions called Fricke eigenforms at prime power level, just knowing a single vanishing order at one special cusp is enough to identify the function uniquely. This is a strong structural result showing these functions are far more constrained than one might expect.
The second part of the paper addresses the question of which half-integer weights can actually be achieved by irreducible cuspidal eta quotients, and introduces a concept called the Mersmann level. This is the smallest level at which a primitive irreducible cuspidal eta quotient of a given weight exists. The paper proves that every half-integer weight is achievable, and establishes a divisibility condition on these minimal levels, showing that the Mersmann level for weight k/2 must be divisible by a rapidly growing power of 2. This connects to classical work by Zagier and Mersmann cataloguing fundamental eta quotients of low weight.