1 VAGRANT, MENIAL, AND ARTISAN CASTES. 335
2 figures for the Dehli and Hissar divisions and for Ambala referring to him and
3 not to the Koli of the hills. The former is probably nothing more than a
4 Chamar tribe immigrant from Hindustan ; the latter, of Kolian orioin. The
5 two would appear, from Mr. Anderson^s remarks quoted in section 6.50, to
6 meet in the Si« aliks. General Cunningham believes that the hills of the
7 Panjab were once oeciipied by a true Kolian race belonging to the same group
8 as the Kols of Central India and Behar, and that the present Kolis are very
9 probably their representatives. He points out that dd the Kolian for water
10 is still used for many of the smaller streams of the Simla hills, and that there
11 is a line of tribes of Kolian origin extending from Jabbalpur at least as far
12 as Allahabad, all of which use many identical words in their vocabularies, and
13 have a common tradition of a hereditary connection with working in iron.
14 The name of Kiilu, however, he identifies with Kulinda, and thinks that it
15 has nothing in common with Kol. Unfortunately Kola is the ordinary name
16 for any inliabitaut of Kulu ; and though it is a distinct word from Koli and
17 with a distinct meaning, yet its plui-al Kole cannot be discriminated from
18 Koli when written in the Persian character; and it is just possible that our
19 figures may include some few persons who are Kole, but not Koli.
20 The names Koli, Dagi, and Chanal seem to be used to denote almost all
21 the low castes in the hills. In the median ranges, sucli as those of Kangra
22 proper, the Koli and Chanal are of higher status than the Dagi, and not very
23 much lower than tlie Kanet and Ghirath or lowest cultivating castes ; and
24 perhaps the Koli may be said to occupy a somewhat superior position to, and
25 the Chanal very much the same position as the Chamar in the plains, while
26 the Dagi corresponds more nearly with the Chuhra. In Kiilu the three words
27 seem to be used almost indifferently, and to include not only the lowest castes,
28 but also members of those castes who have adopted the pursuits of respectable
29 artisans. The very interesting quotations from JMessrs. Lyall and Anderson
30 in sections 6.50, 651 give full details on the subject. Even in Kangra the
31 distinction appears doubtful. Mv. Lyall quotes a tradition which assigns a
32 common origin, from the marriage of a demi-god to the daughter of a Kiilu
33 demon, to the Kanets and Dtigis of Kiilu, the latter having become separate
34 owing to their ancestor, who married a Tibetan woman, having taken to
35 eating the flesh of the Yak, which, as a sort of ox, is sacred to Hindus ; and
36 he thinks that the story may point to a mixed Mughal and Hindu descent for
37 both castes. Again he writes : " The Koli class is pretty numerous in
38 ■' Rajgiri on the north-east side of parganah Hamirpur ; like the Kanet it
39 ■' belongs to the country to the cast of Kangra proper. I believe this class is
40 " treated as outcast by other Hindus in Rajgiri, though not so in Bilaspur
41 "and other countries to the east. The class has several times attempted to
42 " get the Katoch Raja to remove the ban, but the negotiations have fallen
43 " through because the bribe offered was not sufficient. Among outcasts the
44 " Chamars are, as usual, the most numerous." Of j^ai'ganah Kangra he
45 v^rites : " The Dagis have been entered as second-class Gaddis, but they
46 " properly belong to a different nationality, and bear the same relation to the
47 •' Kanets of Bangahal that the Sepis, Badis, and Halis (also classed as second-
48 '' class Gaddis) do to the first-class Gaddis." So that it would appear that
49 Dagis are more common in Kangra proper, and Kolis to the east of the
50 valley ; and that the latter aie outcast while the former claim kinship with
51 the Kanet. It will be oliserved tlmt, while Chamars are returned in great
52 numbers from Kangra and the Hill States, Chuhras seem to be included under