Another matter that had to be considered under its different aspects in the course of this work was the treatment of many modern names that are formed according to different laws in different provinces, and in direct contrast to European usage. Family names in our sense of the word do not exist in India. The tendency to establish them here and there has become observable only in our own days, and is still far from being widely imitated; for our purposes this should accordingly be left out of account. The consequence is that each component part of a name is of equal importance in our eyes, and names have to be taken in their totality and in the same order in which they stand, with the exception, of course, of titles of honour prefixed to a name. We, therefore, have no choice but to put the "Rev. K. M. Banerjea" and "S. P. Pandit" under "Krishṇamohana Vandyopadhyaya” and “Şankara Páṇḍuranga Pandit."

While adhering strictly to this broad principle for the better guidance of those who would consult this Catalogue, we are not unaware of the particular significance, in many instances, of the separate parts of a compound name. In the following outline of the most salient features in Hindu name-giving, as far as they come under our notice for the present purpose, we classify names according to the different nationalities in each of which a peculiar habit seems to prevail. The nearest to the practice of older Sanskrit authors and almost identical with it is that of the HINDI Writing community, who limit themselves, as a rule, to the juxtaposition of two proper names of deities or of one such name coupled with some well-known attribute. BENGALI names enlarge upon this custom by usually adding in the third place a word denoting caste or dignity, such as Şarman, Deva, Mitra, Dása, Datta, etc. Amongst the Brahmins it is usual to drop the title "Şarman," and to substitute one or two learned degrees gained by them for literary distinction, although from what source or by whose authority may not appear. As they sign themselves usually "Sarmá" in their prefaces, we have inserted this attribute in such names, indicating its optional use, however, by including it in parenthesis. MARATHI and GUJARATI names show a tendency to follow Persian custom in combining with the name of the individual in the first place that of the father in the second, and often subjoining a sort of "nisbeh" denoting residence at a place or descent from a family, formed by an archaic locative such as Parvate, Gháre, or by the derivative -kara, such as Bhaṭavaḍekar or Talekar. In TELUGU and TAMIL names the order of things is reversed, and where there is a family name it precedes the individual name. But so little importance is there attached to it that it is mostly abbreviated into one syllable or entirely dropped, and is sometimes remembered with difficulty, as it is alleged, by the bearer thereof who happens to appear before an English court of law. Keeping this peculiarity in mind, we have sometimes restored the abbreviation where we had the means of doing so by comparing it with the full quotation elsewhere, and where this could not be done (as, for instance, in the case of T. Kristnasawmy Iyer) we have put the name under the first initial in order to remember it when on some future occasion the full name might be met with. That such first names are hereditary is demonstrated by the fact that they are borne in common, by brothers for instance, whereas in the north no common appellation reminds one of a common descent of individuals from one and the same stock.

Now, although we recognise the subordinate position of one part or the other in Hindu names inasmuch as they only serve to distinguish persons of the same common birth-name, it has been found impracticable to mark this position outwardly, in smaller type, as we occasionally do when assigning mere epithets to individuals in English terms. It would have appeared less pleasing to

the eye in print and besides would, in some degree, have confused the alphabetical arrangement of the headings. Clearly, the simplest rule for treating Hindu names is therefore to take them as they stand, without making any difference between nomen, praenomen, cognomen or agnomen. This ought always to be kept in mind when, for the identification of names, booksellers' catalogues happen to be compared with ours, their practice not being necessarily circumscribed by the observation of analogous cases and of settled rules derived therefrom.

The only case where we are bound to draw a distinction between a surname and a Christian name occurs amongst SINHALESE names, where it is the custom to prefix to one or two compound buddhistic names a Portuguese Christian name (e.g. Don Andris da Silva), besides adding the name of a place in the locative in -e (e.g. Baṭuvantuḍávé, Waskaduwe) at the end as in Marathi.

As regards the arrangement of titles under the same heading, the principal consideration has been to illustrate their meanings as much as possible by their relative positions to each other. This does not apply to cross-references, which always precede the main-titles, and follow each other in the alphabetical order of the main-headings to which they refer. In the latter, the editions of the texts are grouped together in chronological order, so far as is compatible with general considerations of usefulness. Translations in European languages without the texts are always placed after the texts themselves, and whenever a doubt exists, the place of such works at the beginning or end of the heading will help to determine their nature. The language of the title-page in such a case determines the language of the whole book, an inference that is not to be drawn with regard to title-pages in any modern Indian language. In such event the text must always be understood to be Sanskrit (or Pali), whatever dialect the title-page happens to be composed in. Mere translations of Sanskrit works in these languages are excluded here, being reserved for special catalogues.

In conclusion, it may be well to advert once more generally to the difficulties that had to be contended with in transforming what had been a catalogue containing detailed descriptions of books, likely to be expanded day by day by fresh insertions and to have titles displaced accordingly,

Bibliographical Manual, free from useless reiterations, and yet clear in its conciseness. The new shape it assumes in print often suggests, of itself, new modes of practice, the consistent carrying out of which calls for the greatest attention in a work so interwoven with cross-references. The judicious shortening in one place, the useful enlargement in another, the necessity of leaving other people's mistakes untouched without letting one's eyes get blinded to one's own anomalies, together with a number of other minutiae, offered frequent chances of error, and called for constant watchfulness in the revision of proofs. That errors should, under these circumstances, have crept in where least suspected, it is but natural to assume; after testing again and again, however, none of a serious kind have come to light. Those that were discovered are rather in the nature of excusable oversights affecting in nowise the general disposition of the work.

The one point least satisfactory to the Editor himself is the description of the sizes of books. Every librarian knows that in this respect it is difficult enough to lay down a general rule for European books. Counting so many leaves to a sheet may seem to afford the safest standard for putting down a book as a 4to. or an 8vo. But as sheets differ considerably in size, so must the leaves into which they are folded, and books most incongruous in size would, by adopting such a rule, often fall under the same category. The inconvenience of a practice like this becomes still more apparent when we try to apply it

to books printed transversely on oblong leaves not corresponding in shape to anything we are accustomed to in Europe. Some of the inconsistencies accruing from this circumstance have been removed in the course of the final revision, others, it is feared, may have been left remaining. But until a general understanding on this point has been arrived at, it really does not greatly matter which way the decision leans in individual cases. The information regarding the pagination of books of non-European origin, inserted just before going to press, will, it is hoped, remedy in some degree the uncertainty arising from the absence of a well-defined law on the description of sizes of books, nay, act as a much safer guide in the expectation one often forms about a book with a grand-sounding title which afterwards proves to be moderate in dimension and trivial in matter.

The foregoing remarks may, at first sight, appear somewhat lengthy, but by discussing those topics once for all, much useless repetition in the Catalogue itself has been avoided, and some useful hints have been afforded, it is hoped, to those who, when engaged in similar work, want to compare for the purpose of identifying or distinguishing Sanskrit books coming under their notice with the descriptions of such as are contained in the Collection of the British Museum.

BRITISH MUSEUM,

May, 1876.

ERNST HAAS.

ACADEMIES, ETC.-AGHORANATHA.