rise of the lake’s level by the melting of the snows, which often coincides with Lent.  I may add that Moorcroft was told respecting a sacred pond near Sir-i-Chashma, on the road from Kabul to Bamian, that the fish in the pond were not allowed to be touched, but that they were accustomed to desert it for the rivulet that ran through the valley regularly every year on the day of the vernal equinox, and it was then lawful to catch them.

Like circumstances would produce the same effect in a variety of lakes, and I have not been able to identify the convent of St. Leonard’s.  Indeed Leonard (Sant Lienard, G. T.) seems no likely name for an Armenian Saint; and the patroness of the convent (as she is of many others in that country) was perhaps Saint Nina, an eminent personage in the Armenian Church, whose tomb is still a place of pilgrimage; or possibly St. Helena, for I see that the Russian maps show a place called Elenovka on the shores of Lake Sevan, N.E. of Erivan.  Ramusio’s text, moreover, says that the lake was four days in compass, and this description will apply, I believe, to none but the lake just named.  This is, according to Monteith, 47 miles in length and 21 miles in breadth, and as far as I can make out he travelled round it in three very long marches.  Convents and churches on its shores are numerous, and a very ancient one occupies an island on the lake.  The lake is noted for its fish, especially magnificent trout.

(Tavern. Bk.  III. ch. iii.; J.  R. G. S. X. 897; Pereg.  Quat. p. 179; Khanikoff, 15; Moorcroft, II. 382; J.  R. G. S. III. 40 seqq.)

Ramusio has:  “In this province there is a fine city called TIFLIS, and round about it are many castles and walled villages.  It is inhabited by Christians, Armenians, Georgians, and some Saracens and Jews, but not many.”

NOTE 7.—­The name assigned by Marco to the Caspian, “Mer de Gheluchelan” or “Ghelachelan,” has puzzled commentators.  I have no doubt that the interpretation adopted above is the correct one.  I suppose that Marco said that the sea was called “La Mer de Ghel ou (de) Ghelan,” a name taken from the districts of the ancient Gelae on its south-western shores, called indifferently Gil or Gilan, just as many other regions of Asia have like duplicate titles (singular and plural), arising, I suppose, from the change of a gentile into a local name.  Such are Lar, Laran, Khutl, Khutlan, etc., a class to which Badakhshan, Wakhan, Shaghnan, Mungan, Chag-hanian, possibly Bamian, and many others have formerly belonged, as the adjectives in some cases surviving, Badakhshi, Shaghni, Wakhi, etc., show[2] The change exemplified in the induration of these gentile plurals into local singulars is everywhere traced in the passage from earlier to later geography.  The old Indian geographical lists,