nildenmi lujvo

x1 [possibly dimensionful number] is the local density/concentration of/in x2 (object) per unit x3 at locus x4, of density type x5; x2 measures x1 units of x5 per unit x3 at locus x4.

x5 might typically refer to count - not mass. Not necessarily (although possible) volumetric mass density; could be temporal, per unit area, per unit length, or other (x3); could be non-mass-based (such as for electric charge or probability densities; x5); relative concentration would be a count density (x5) per count of another substance (x3). Can be applied to number/count; thus, the units may vary and in fact the quantity might even be unitless. In a sense, intensities and fluxes are spacial-temporal densities. It is recommended that x2 be a physical entity/object/field or probability distribution in some sense (not merely have physical quality/meaning/value); thus, while this word could be applied to many things described with the word "per", it would probably not make much sense to call velocity a temporal displacement density (although in some contexts it might). Notice that "local" implies derivatives (Calculus differentiation).

In notes:

x1 (unitless/dimensionless number; li) is the specific gravity (density or concentration ratio) of x2 (substance) with respect to/by standard/per unit of density/concentration of x3 (substance), where density/concentration of each substance is being taken per unit x4 = nildenmi3 at locus x5 = nildenmi4 of density/concentration type x6 = nildenmi5.
x1 is the vector flux (flow) of quantity/substance (count) x2 through (geometric/imaginary) hypersurface/embedded manifold x3 per unit hyperarea per unit time.
x1 is the scalar flux (total flow) of quantity/substance x2 through bounding Jordan orientable manifold x3 under conditions and conventions of definition x4 (including actual orientation of x3, signum of outflux, etc.)