Curious reflection.—­Observed in ghee used as lamp-oil, a bubble ascending from the surface of the water on which it floated, met by another descending; the deception of this is perfect.  That it is due to reflection, is apparent from the variation of the length of the descent, according to the angle under which it is viewed.  When viewed from beneath at a very oblique angle, the descent is complete, but if viewed parallel to the surface, no appearance of the sort occurs.  The reflection is due to the surface of the ghee which appears to be more dense than the rest, probably more oily; this mathematical reflection may suggest others of a moral nature, touching our liability to mistaken views of things, from observing only one side.

Old Candahar is about three miles to west of the new town; it is immediately under a steep limestone range, running about southwest, and not exceeding 500 feet in height.  It bears marks of having been fortified, and at either extremity remains of forts are still visible.  The fort of forty steps is at the north end of the range.  The town is in complete ruins; indeed none of the edifices are visible except those that occupy the mound of stones, (with which they are partly built) probably the site of the citadel.  On three sides, the town is fenced by two respectable ditches, the outer one about 50 yards wide; both are now, especially the outer, beds of marshes; they were supplied by cuts from the Arghandab river.  Wells exist however.  There is one white mosque in good preservation.  The works were strong, and much better than the very indifferent ones of new Candahar; and the walls of the town were prolonged up the face of the hills.

About Candahar, conical houses occur, probably for granaries.  A curious mosque cut out of the rock in situ, is seen on the Girishk road, with a flight of steps leading to it, cut in like manner out of the rock.  There is also in the same quarter the fort of Chuhulzeenat, or forty steps; a work not of very considerable extent; and as in other Asiatic countries I have visited, troughs are cut in rocks for separating grain from the husk.  But there is no work to be seen indicating vast labour or any genius.

Some remains of good pottery may be picked up; and the earth of which the works, etc. were made, is filled with remains of coarse pottery.

27th.—­Moved four miles to Shorundab, the country is very barren:  not much Joussa:  the water is brackish at our present encampment, which is within sight of Babawallee.

28th.—­Proceeded to Kileeyazim, ten and a quarter miles, marched at 2 P.M. and reached the place at 6 P.M., the camels arriving one hour afterwards:  the ground is generally good, throughout stony, difficult in places and undulated, particularly in two situations occasioned from cuts.  There is a square fort, situated at the halting place with a tower at each corner, and on north face two; as well as towers at the gate:  but without windows. Joussa is abundant, as also grass along the cuts.  Salsola rotundifolia, a Chenopodia, and a curious prickly, leafless Composita and Joussa occur, the latter most common, Artemisiae sp.  Also rock pigeons and the raven.  Halted one mile to the east of the fort.