NOTWITHSTANDING the voluminous literature on the subject of the liquefaction of air, I fail to find what seems to me a really satisfactory exposition of its theoretical features based on fundamental thermodynmnic facts and laws. Some important theoretical points are either unnoticed or only incidentally touched on,--e.g., the fundamental influence of molecular attractions on the possibility of liquefaction at all; and, more particularly, the peculiar r61e molecular attractions play in a continuous isothermal process, where all the heat abstracted from the compressed air is confined to what is given out at readily available temperatures such as is afforded by the ordinary water supply. For instance, in Poynting and Thomson's Heat, 1 the authors of which may be regarded as occupying positions of the highest authority on any subject to which they give their attention, the cooling in such process is merely incidentally attributed to the work done in overcoming molecular attractions; and the reasons for the statement are given only by calling attention to an apparent analogy of the process to that of the porous plug experiment. From this fact I conclude that probably no more satisfactory, concise, and non-mathematical explanation was known to them or it would have been given, or at least referred to. A more complete theoretical explanation may have been published, but, if so, I also am unacquainted with it. I know of no more satisfactory, concise explanation of the theory than they offer; but, in my judgment, their exposition is too brief and incomplete to enable one to comprehend the process as a whole. Other writers, again, attribute the cooling to work done on the gas b)r itself. While this statement is true, in a sense, it may be misleading, because in the continuous process work is continually * Communicated by the author.