1 86 PANJAB CASTES.
2 whlct they were tabulated and the final figures compiled will he explained in Boot II under the
3 head? Tabulation and Compilation. The whole process was intended to be merely a rough one.
4 The detailed tabulation is now in progress, and I hope wi'.hiu the next few years to bring out
5 detailed tables of tribes and clans for the whole Panjab. But besides inaccuracies that will have
6 crept into the work of tabulation, there are several causes of error inherent in the material. In
7 the first place the spelling of local names of tribes, as rendered by the enumerating staff, varied
8 extraordinarily. Some were evidently mere variations, as Dhaiiwal, Dhah'wal and Dhaniwal ;
9 some I liuew to represent separate tribes, as Sidhu and Sindhu, Chhiua and Chima ; some I am
10 still in donbt about, as Euta and Bhiitta, Sara and Sarai. In working with a staff not always
11 acquainted with the names of the clans, figures referring to two different tribes must often
12 have been joined together, and other tigui'es wTongly omitted because of some variation
13 in the spelling. Another source of error doubtless was the uncertainty regarding tlie
14 woman's clans discussed in section 334. On a future occa-ion I would tabixlate sub-
15 divisions of castes for males only. Again many of the people are presented twice over
16 in two columns. Thus the Sial are Punwar Rajputs by origin. Suppose that 1,000 Sials
17 returned themselves as Rajput Punwar Sial, another thousand as Sial Punwar, another 1,000
18 as Rajput Sial, and a fourth 1,000 as Rajput Punwar. All the 4,000 people would be shown in
19 Table VIII as Rajput ; but in the details of tribes we should have 3,000 Sial and 3,000
20 Punwar or 6,000 in all. This was quite unavoidable so long as only one tribal division was
21 tabulated ; but as a fact the cases in which this happened were few, or at least the numbers
22 affected small. I had all cases in which the same people were entered twice over shown in a
23 separate memorandum attached to the tribes table, and wherever the numbers were at all con-
24 siderable I have mentioned the fact of their double inclusion in the text. This double entry
25 occurred most often with the Jat tribes, who, in order to fill up their three columns, entered the
26 Eajput tribe from which they claimed origin as well as their omi Jat tribe, so that we bad
27 people returning tlicmselves as Jat Sidhu Bhatti, and such people appear among the Jat tril)es
28 both as Sidhu and as Bhatti.
29 370. Proposals for next Census. — What then is best to he done at next Census ? It will
30 be seen that mauy of the difficulties are due to the intrinsic difficulty of the question and to the
31 varying nature of caste in the Panjab. So far as this is tlie case no scheme will help us.
32 In one respect, however, I hope that the task will be made much easier by next Census. I hope
33 by then to have brought out classified lists of all the tribes and clans returned in the present Census.
34 The way in which they will facilitate the treatment of the subject is explained in the section on
35 Tabulation. If I liadhad such a classified list my task on this occasion would have been easy enough ;
36 and it is I think one of the most valuable results of the present Census that it has given us
37 materials for the preparation of such a list. With such a list the three columns of the schedule of
38 1881 are almost perfect in theory. But 1 do not think they worked as well in practice. I believe
39 that the three columns which they erroneously thought they were hound to fill up, puzzled
40 both people and staff, and caused a good many of our difficulties. Thus in future I would have
41 but two columns, and would head them Quam and Shdich. I would not care whether caste or
42 tribe was entered in the first column, as the classified list would show the tabulator how to class
43 the tribe ; and I would hope that the second column at any rate would generally give tribe.
44 In very many cases it would not. There would be entries like Biloch Rind instead of BUoch
45 Laghari, Brahman Bashisht instead of Brahman Sarsiit, Banya Kasib instead of Banya Agarwal,
46 and so forth. But on the whole I think it would be better to accept the fact that the entries
47 must be incomplete, whatever scheme be adopted ; and would pi-efer the certainty of error of the
48 two columns, rather ttan the confusion and perplexity which the tteee columns cause to those
49 concerned in the enumeration. Above all things I would avoid tlie words asl and got. I would
50 let the patwaris, who should make the preliminary record, exercise their discretion about entering
51 high castes for menials or artisans, du-ecting them to show the caste by which the people were
52 commonly known in the village. I would tabulate both males and females for tribes and clans,
53 and arrange them in order of numbers ; and 1 would have the Deputy Superintendent personally
54 examine the tribal tables for all above say 500, before compiling bis final caste tables. Such
55 an examination would do an immense deal towards increasing the accuracy of tlie caste figures ;
56 but it was impossible in the present Census owing to the double sub-division. I would show in
57 my tribal tables tlie figures for males only, though those for females must be tabulated in the
58 first instance in order to allow of transfer of entries from one caste heading to another.
59 371. Bibliography. — The most detailed and accurate information available in print regard-
60 ing certain, and those the most important from an administrative point of view, of the Panjab
61 castes is to be found in the numerous Settlement Reports, and more especially in those of recent
62 years. Unfortunately they deal almost exclusively with the landowning and cultivating castes.
63 Sir H. EUiott's-vfiaoes of the N. W. P., edited by Mr Beames, is, so far as it goes, a mine of
64 Information re^rding the castes of the eastern districts. Sherring's Hindu, Castes contains
65 much information of a sort, the first volume being really valuable, but the second and third being
66 infinitely less so j while the whole is rendered much less useful tliau it might be by the absence
67 of any index save one that maddens the anxious inquirer. On the ancient form of the institution