Scientific discoveries often necessarily involve observations at the ragged edge of invisibility. These were indeed through a glass darkly. Through the unwanted impediment of a badly scratched and cleaned airplane window!

Tomas is looking downwards below the horizon. All the halos are subhorizon arcs. The long arc is the subparhelic circle. At left, other arcs cross it at the subanthelic (antisolar) point. All these are rare sights but the extreme rarity is the white spot at right – a sub 120° parhelion.

Are we sure? The montage images are from a lens at a constant known zoom setting and taken within 17 seconds of each other. They are aligned on the reasonably sharp end of the bright part of the subparhelic circle. The estimated angular distance from the subanthelic point to the right hand bright spot is within a 5° of the required sub 120° parhelion position. The second montage at page bottom is from two more images 13 seconds apart and gives the same result.

The above horizon 120° parhelion is relatively rare partly because unlike its 22° counterpart it is colourless and therefore are less easy to see against a cloud background. Rays forming it (left) follow a complicated path through plate crystals drifting with their large hexagonal faces nearly horizontal. Rays that undergo an additional internal reflection from the lower crystal face form the elusive sub 120° parhelion.

The wayward ray path is best achieved if the plates are thicker than is usual. It is also helped if the crystals are not regular hexagons but have at least one set of alternate long and short side faces.

The intensity variation along the parhelic circle gives some clues to what is happening. The top HaloSim ray tracing used plate crystals with alternate side face lengths of 4:1. The intensity along the bright part of the subparhelic circle is fairly uniform as it is in the images.

The lower ray tracing used thick regular hexagons and they produced an uneven subparhelic intensity with pronounced Liljequist parhelia brightenings and further unobserved intensity near the sub 120° parhelion itself. But there are many parameters including the crystal tilts and these few ray tracings are only indications.

The diffuse arcs crossing the subanthelion were produced in the simulation by adding singly oriented column crystals. Parry columns might also have contributed to those in Tomas’s images and to his subparhelic circle.

Whatever the finer details of their origin, the sub 120° parhelion images are impressive and probably a first. Photos at night using a lamp to illuminate crystals nucleated by ski-slope snow machines have caught the subparhelion before but not as here with the sun shining down on crystals naturally formed in the clouds.