Günther Konnen has drawn attention to this famous image showing disconnected rainbows. J Dijkema imaged it in the Pacific Ocean.
The upper primary bow (with a fainter supernumerary) was formed by falling rain.
The lower bow was made by drops of seawater thrown up by waves against the ship’s side. The seawater bow has a slightly smaller radius (by about 0.8°). The difference would not normally be apparent but here it is obvious by comparison with the rain water bow.
Seawater, because it contains dissolved sodium, calcium and magnesium salts, is slightly denser than pure water and also has a greater refractive index. The lower diagram shows two minimum deviation rays going through spheres of different refractive index to form primary bows. Seawater and rainwater are too alike to show clearly distinguishable rays and so instead, ray paths are shown for a water sphere (n = 1.33) and a glass sphere (n=1.51).
As the refractive index increases, the incoming ray that forms the primary bow (minimum deviation) moves inwards to so that if it were undeviated it would pass closer to the drop centre. At a sufficiently large refractive index the ray actually passes through the centre, the deviation angle approaches 180°, and there is no longer a rainbow. Highly refractive substances cannot form rainbows.