In Macroscopic approach of thermodynamics the substance is considered to be continuous whereas
every matter actually comprises of myriads of molecules with intermolecular spacing amongst them.
For analyzing a substance in aggregate it shall be desired to use laws of motion for individual molecules
and study at molecular level be put together statistically to get the influence upon aggregate. In statistical
thermodynamics this microscopic approach is followed, although it is often too cumbersome for practical
calculations.
In engineering thermodynamics where focus lies upon the gross behaviour of the system and
substance in it, the statistical approach is to be kept aside and classical thermodynamics approach be
followed. In classical thermodynamics, for analysis the atomic structure of substance is considered to
be continuous. For facilitating the analysis this concept of continuum is used in which the substance is
treated free from any kind of discontinuity. As this is an assumed state of continuum in substance so the
order of analysis or scale of analysis becomes very important. Thus, in case the scale of analysis is large
enough and the discontinuities are of the order of intermolecular spacing or mean free path then due to
relative order of discontinuity being negligible it may be treated continuous.
In the situations when scale of analysis is too small such that even the intermolecular spacing or
mean free path are not negligible i.e. the mean free path is of comparable size with smallest significant
dimension in analysis then it can not be considered continuous and the microscopic approach for
analysis should be followed. For example, whenever one deals with highly rarefied gases such as in
rocket flight at very high altitudes or electron tubes, the concept of continuum of classical thermodynamics
should be dropped and statistical thermodynamics using microscopic approach should be followed.
Thus, in general it can be said that the assumption of continuum is well suited for macroscopic approach
where discontinuity at molecular level can be easily ignored as the scale of analysis is quite large. The
concept of continuum is thus a convenient fiction which remains valid for most of engineering problems
where only macroscopic or phenomenological informations are desired.