148 Thi:ry of Horizontal RefraBkn. — Ligkt-Houftj, 



refracted through the two mediums; fome parts of the hull appearing cre£l, and fomff 

 inverted. Suppofe the dotted line //, Fig. 4. an indefinite pencil of rays pafling from be- 

 tween the inverted and credl parts of the cbjetl, or the upper part of the hull of the 

 veflel to the eye (for the lower part of the hull could not be obfervedl, the objects cannot 

 appear inverted, except the angles at the eye a Ac and a .\ d exceed the angle a A i ,■ for 

 the intermediate fpace could only be contracted by the fecondary pencil of rays. The 

 lengths of the inverted, compared with the eredt image of the fail, is as the fines of ihc 

 angles at the eye a A < to /' A ti ; and the angle at the eye a Ad, made by tlie two pcncik 

 of rays from the fame point near the head of .the fail, mud be double the angle a A i, 

 when the inverted image is as tall as the ere£t. In this cafe the fines of the angles a At, 

 a Ac, a Ad, Fig. 4, are proportional to the altitudes ab, ac, a d, in the magnified view of 

 the vefTel Fig. 3. 



Under this confideratlon, no inverted image of the fail will be formed, until the angle 

 at the eye, made by the two refratlcd pencils of rays na and dd, exceed the angle made by 

 ma and b i, the apparent height of the fail of the veflel ; for, were thofe angles equal, the 

 inverted fail would only be contracled into the parallel of altitude of the boom b, ajid render 

 the appearance confufed, as in the hull of the velTcl. 



Refpefting the exiftence of two pencils of rays entering the eye from every point of an' 

 objc£l not more elevated th.ui a, or lefs than i. Fig. 3, in this (late of the atmofphere, I 

 cannot bring a (tronger proof than that of the ftrength of a light when the rays pafs near 

 the horizon of the fea, proved by the following obfervations : 



Going down Channel about five years ago in the Trinity yacht, with fcveral of the 

 elder brethren, to infpedt the light-houfes, &c. I was told by fome of the gentlemen 

 who had" been on a former furvey, that the lower light of Portland was not fo Urong 

 as the upper light at near diftances ; but that at greater diftances it was much ftronger. 

 I fufpecled that this difference arofe from the lower light being at or near the horizon of 

 the fea, and mentioned it at the time ; but afterwards had a good opportunity of making the 

 obfervation. We paffed the Bill of Portland in the evening, fleering towards the Start; a 

 frefli breeze from the northward, and clear night. \Vhen we had run about five leagues 

 from the lights, during which time the upper light w-as uuiverfally allowed to be the ftronger, 

 (feveral gentlemen keeping watch to make obfervations thereon), the lower light drawing 

 near the horizon fuddenly fhone with double lullre. Mr. Strachan, whofe fight is weak, had 

 for fome time before loft fight of both lights ; but could then clearly perceive the lower 

 light. I then went aloft (as well as others), but before I got half maft up, the lower light 

 was weaker than the upper one. On coming down upon deck, I found it again as ftrong as 

 before. We proceeded on, and foon loft the lower light from the deck ; and upon draw- 

 ing the upper light near the horizon, it, like the former, ftione exceeding bright. I again 

 went aloft, when it diminiflied in brightnefs ; but from the maft-head I could then fee the 

 lower light near the horizon as ftrong as before. This is in confequence of the double 

 quantity of light entering the eye by the two pencils of rays from every point. To 

 illuftrate which, we compare the veflel Fig. 4. to a light-houfe built upon the (hore, 

 and A the place of the obfervcr ; and having brought down the light fo low as to view it 

 in the direflion a a, another light would appear in the horizon at x, from the pencil dd\ 

 and had the veflel been ftill enough to have obfcrved it at this time with a good glafs, I 

 doubt not but the two images might have been diftiiidly feen. As the light dropped (by 



increafing 



