1 CASTE IN THE PAN JAB. 85
2 there, to be aiTanged afterwards. But wheu we came to examiue tlic sebedales we found that
3 the separate entries in the caste column alone were nnmhered bv thousands, while the sub-
4 divisions were numbered by tens of thousands. I certainly had not, and I do not believe that any
5 body else concern&i had, the very faintest conception of how numerous the entries would be. At
6 any rate it was obviously quite out of the question to tibulate and examine them aU before com-
7 pilation ; and what was done was to deal with the entries in the first or caste column only, so far
8 as the Ci epilation of the final Census Table VIII was concerned. Even those entries I was com-
9 I.elled, for reasons given in the Chapter on Tabulation, to allow the Divisional Officers to classify
10 for themselves where there appeared to he no reasonable doubt as to the classification. With the
11 headings for which they rcturne.1 separate figures I dealt as is described in the Chapter on Com-
12 pilation. The figures for the sub-divisional entries were tsbulated in detail ; but only certain
13 selected entries were taken out to be used in the Census Report, the principles on which the selection
14 was made being explained in the Chapter on Compilation.
15 368. Nature and degree of error in the final figures. —Thus the figures as now given in
16 the abstracts and appendices of this report are liable td error in several ways. In the first place
17 many members of a caste or tribe entered as their caste some race to « hich they arc pleased to refer
18 their origin in remote antiquity. For instance, some Gakkhars returned themselves as Gakkhar and
19 others as Mughal, and are shown under those headings respectively in the final tables, which there-
20 fore do not give the tofcU number of Gakkhars in the Panj.'ib. So some low caste men returned their
21 caste as Rajput or Mughal or Quvesh ' out of joke ' a< several Deputy Commissioners note. On the
22 other hand ?ome men of good ca-f e, ?ucb as Sial, Khokhar, or Mughal, who were following the
23 trade of weaver or carpenter, returned their caste a- Paoli or Tarkhan, though the adoption of that
24 hereditary occupation liad been in many ease; too recent to ha\ c brought about a change of ca»te.
25 This la~t error was for the mo-t part confined to the Western Plains. Again, persons who
26 belonged to the same tribe and returned that tribe as their caste will have been differently
27 claj=ed in different divisional offices, or classed under one heading in one division and returned
28 separately and then classed by myself und'T another heading in another division. Thus the
29 Phaltis will have been cla=--ed as Jats by the Dcrajat and as Rajputs by the Rawalpindi office.
30 S.) the Langahs were classed as Jats in Mult.in, while the Dcrajat returned them separately and
31 I classed them as Pathans. These errors however affect only those cases where the tribe was
32 returned and not the caste. Where a man returned himself as Jat, Rajput, Pathan and so fc^rth,
33 he was treated as such although the tribe he gave might raise suspicion as to the correctness of
34 the returns. Moreover the errors, if they must be so called, do represent actual facts. The
35 Bhatti is a Rajput in Rawalpindi because there Rajputs are recognised. In the Derajat he is
36 a Jat, because there no distinction is drawn between Jat and Rajput. And it must be remembered
37 that though the cases in which the errors detailed above occurred are numerous, the total figures
38 affected are seldom large. There were certainly hundreds, I believe there were thousands of
39 so-called castes returned in the Multaii division which only included ten or fifteen people in the
40 whole division. The great ma-s of each caste returned themselves rightly and are shown correctly
41 in our tables: the items that are wrongly classed are wholly insignificant in their total amount as
42 compared with the items that are rightly cla=sed. But there are exceptions to this statement.
43 The distinction between Jat and Rajput is so indefinite and so variable that it can hardly be
44 called a mistake to class a tribe as Jat in one place and Rajput in another. This however has
45 been done. But I have picked out the figiires in each case and put them side by side in the
46 abstracts contained in the section on these two castes, and I think the error which has not been
47 corrected may be taken as exceedirgly -mall. It is now in each man's power to transfer the
48 figures for any tribe from Jat to Rajput or vice artd, according to iudividnal taste. The other
49 chief exceitious are in the case of Mughals and Shekhs. For Sliekhs I was prepared. I knew
50 that all sorts of low caste men, recent converts to Islam, would retiu-n themselves as Shekh ;
51 and I had the figures examined with a view to separate these, and the details will be found in
52 the te.xt of this chapter. But I did not know that in some parts of the we-tern Panjab Mughal
53 was as favour'd: .' .■•■; ; i 'I'.ixus origin as Shekh is in other parts of the Province, and 1 have not
54 had the details worked out so carefully. Still almost all the large numbers have been separated
55 from these two entries. So with Pathans'. Many people, such as Dilazak, have returned them-
56 selves as Pathans who do not really belong to the race ; but their claim to the name is often admitted,
57 and they have become in a way affiliated to the nation. Thus the considerable ciTors in the caste
58 tables, as corrected in this chapter, amount to this ; that there is a confusion between Jat and Eaj-
59 piit and between Pathan and certain allied races, which exists in actual fact fully as much as in the
60 figures ; that some tribes or castes have been wrongly shown as Mughal and Shekh ; and that some
61 oi the arti-an castes have been shown as belonging to the higher castes, while some of the higher
62 Civstes have been included in the artisan castes merely because they followed their occupation.
63 Taking the Province as a whole the errors are probably insignificant, and hardly affect the
64 general distribution of the population by caste. They are probably greatest in tlji cis-Indus Salt-
65 range tract, where the tendency to claim Mughal origin m strongest.
66 369. Error in the figures for tribes and sub-divisiODS.— The figur^ for tribes and
67 sub-divisions given in this chapter are professedly only rough approximatiaibs. The manner
68 d2