diminution and arrest of growth shows that the units have fallen into such relative positions that re-distribution is no longer so facile. When, therefore, we see that gamogenesis recurs only when growth is decreasing, or has come to an end, we must say that it recurs only when the organic units are approximating to equilibrium--only when their mutual restraints prevent them from readily changing their arrangements in obedience to incident forces. That units of like forms can be built up into a more stable aggregate than units of slightly unlike forms, is tolerably manifest _à priori_. And we have facts which prove that mixing allied but somewhat different units, _does_ lead to comparative instability. Most metallic alloys exemplify this truth. Common solder, which is a mixture of lead and tin, melts at a much lower temperature than either lead or tin. The compound of lead, tin, and bismuth, called "fusible metal," becomes fluid at the temperature of boiling water; while the temperatures at which lead, tin, and bismuth become fluid are, respectively, 612°, 442°, and 497° F. Still more remarkable is the illustration furnished by potassium and sodium. These metals are very near akin in all respects--in their specific gravities, their atomic weights, their chemical affinities, and the properties of their compounds. That is to say, all the evidences unite to show that their units, though not identical, have a close resemblance. What now happens when they are mixed? Potassium alone melts at 136°, sodium alone melts at 190°, but the alloy of potassium and sodium is liquid at the ordinary temperature of the air. Observe the meaning of these facts, expressed in general terms. The maintenance of a solid form by any group of units implies among them an arrangement so stable that it is not overthrown by the incident forces. Whereas the assumption of a liquid form implies that the incident forces suffice to destroy the arrangement of the units. In the one case the thermal undulations fail to dislocate the parts; while in the other case the parts are so dislocated by the thermal undulations that they fall into total disorder--a disorder admitting of easy re-arrangement into any other order. For the liquid state is a state in which the units become so far free from mutual restraints, that incident forces can change their relative positions very readily. Thus we have reason to conclude that an aggregate of units which, though in the main similar to one another, have minor differences, must be more unstable than an aggregate of homogeneous units. The one will yield to disturbing forces which the other successfully resists.