If I define every thing as a whole made of its parts, these parts should be things as well – but made of what?
If every part is composed by smaller ones, we fall in a regress that leads to infinitesimal elements. If we stop at an indivisible unit, it cannot be composed by parts nor multiples of itself: we can define it only through relations in a different reference system.
i.e: Suppose that the "minimum unit" that composes the forms in the lower part of the image is the upper triangle (let's call it the "elementary triangle"). Unlike these composite forms, you cannot describe the triangle in terms of triangles - it already is - but only on the basis of relations outside the system of the figures-made-by-elementary-triangles. To define an elementary triangle, for example, you can use colors, lines, ink particles, its symbolic value, mathematical formulas and so on. Colour, lines and ink, in fact, are not composed by elementary triangles. The minimum unit of any closed system can only be defined through reference systems where it's not the minimum unit.
As noted by Mauro Allegranza, I have to distinguish the philosophical views about Atomism (mainly ancient), with the related debate about infinitesimals and indivisibility, from the modern atomic physics. According to modern science, material stuff is made of atoms; atoms in turn are made of subatomic particles that presumably have no substructure, i.e. they are not composed of other particles. If they are not divisible as it seems, they can be defined only through relations in a different reference system (i.e interactions with other particles)