Now, because that heat and cold are equally diffus’d every way; and that the further it is spread, the weaker it grows; hence it will follow, that the most part of the under Region of the Air will be made up of several kinds of lentes, some whereof will have the properties of Convex, others of Concave glasses, which, that I may the more intelligibly make out, we will suppose in the eighth Figure of the 37. Scheme, that A represents an ascending vapour, which, by reason of its being somewhat Heterogeneous to the ambient Air, is thereby thrust into a kind of Globular form, not any where terminated, but gradually finished, that is, it is most rarify’d in the middle about A. somewhat more condens’d about BB, more then that about CC; yet further, about DD, almost of the same density with the ambient Air about EE;, and lastly, inclosed with the more dense Air FF, so that from A, to FF, there is a continual increase of density.  The reason of which will be manifest, if we consider the rising vapour to be much warmer then the ambient heavie Air; for by the coldness of the ambient Air, the shell EE will be more refrigerated then DD, and that then CC, which will be yet more then BB, and that more then A; so that from F to A, there is a continual increase of heat, and consequently of rarity; from whence it will necessarily follow, that the Rays of light will be inflected or refracted in it, in the same manner as they would be in a Concave-glase; for the Rays GKI, GKI will be inflected by GKH, GKH, which will easily follow from what I before explained concerning the inflection of the Atmosphere.

On the other side, a descending vapour, or any part of the air included by an ascending vapour, will exhibit the same effects with a Convex lens; for, if we suppose, in the former Figure, the quite contrary constitution to that last describ’d; that is, the ambient Air FF being hotter then any part of that matter within any circle, therefore the coldest part must necessarily be A, as being farthest remov’d from the heat, all the intermediate spaces will be gradually discriminated by the continuall mixture of heat and cold, so that it will be hotter at EE, then DD, in DD then CC, in CC then BB, and in BB then A. From which, a like refraction and condensation will follow, and consequently a lesser or greater refraction, so that every included part will refract more then the including, by which means the Rays, GKI, GKI, coming from a Starr, or some remote Object, are so inflected, that they will again concurr and meet, in the point M. By the interposition therefore of this desending vapour the visible body of the Star, or other Object, is very much augmented, as by the former it was diminished.

From the quick consecutions of these two, one after another, between the Object and your eye, caused by their motion upwards or downwards, proceeding from their levity or gravity, or to the right or left, proceeding from the wind, a Starr may appear, now bigger, now less, then really it would otherwise without them; and this is that property of a Starr, which is commonly call’d twinkling, or scintillation.