above. There are, however, some traces of a diminutive force as in dalotdót dig a little,' and in giligintó 'the (little) golden eyes that form on soup,' properly the little golden ones,' or 'the goldish ones' from gintó 'gold,' which must be compared with the Bisayan formations like tolotigolong 'oldish' and karokabayo 'little horse,' etc., mentioned above.

In a number of cases la, li, lo are infixed in a simple dissyllabic root, no definite modification in meaning being apparent, e. g. :

algkóm 'take, seize as much as the fist will hold,' from the root ákom hold in the hands, contain in the hands,' with infixed la.

bałoki 'double a point in sailing.' Cf. báki 'say something to get from another what one desires' (cf. our "get around" a person), and bóki 'speak in a round-about fashion.' The original idea at the base of all these words was probably 'go around, get around,' with infixed lo.

halálany place something in the way in order that people may not pass through,' from hálang 'some obstacle placed in the way,' with infixed la.

salagip meaning the same as and formed from sagip 'seize, take something which is in the water; adopt the manner of another; hold as a hostage,' with infixed la.

talósok (Supplement') 'be wedged in or stuck fast in the mud (of the leg, or tikin [governing pole of a skiff])' from tások a wedge, or peg,' with infixed lo.

In giligintóthe golden eyes on soup' from the root gintó, we have an example of the same formation as the Bisayan forms with la, li, lo inserted after the first syllable of roots with initial partial reduplication having a diminutive meaning, e. g., tolotigolong 'oldish,' from tigolong 'old.'

The particles la, li, lo in Tagalog are inserted in a number of cases in roots consisting of two identical syllables, a diminutive force being sometimes apparent, e. g. :

balakbák a kind of timber or bark of a tree,' from the root bakbák strip the bark from trees, peel fruit,' with infixed la. dalayray place things in order in such a way that one shall not be above the other' from the root dayray row, file of things,' with infixed la.

salaksák (Supplement') 'cram, ram with a rod, piston, or other instrument,' derived from saksák, which has the same meaning.

salaysay explain,' from the root saysay explain, clear up; place material things in order,' with infixed la.

salayósay blow gently (of the wind),' from the root sayósay with the same meaning, with infixed la.

In some cases in which the vowels of the root are o or i, the first vowel of the derived word or root with infixed particle seems to be changed to a. But this a may be simply the representation of the indistinct vowel in an unaccented syllable. Thus we find:

dalogdóg ‘beat the kettle-drum,' probably from the root dogdóg grind, pound,' with infixed lo and a instead of o in the first syllable.

dalotdót dig a little,' vide dotdót stir with the finger,' with infixed lo.

alignrig shake itself (an animal), shake powder, water, etc., with infixed li. Cf. wagág shake a thing to find out. what is in it,' and wigwig shake, wander; speak in circumlocutions. Since both agirág and wigwig occur,

walignrig may represent a combination of the two, the a in the first syllable being derived from ragrag and the final from rigrig. Similarly

walisuís ‘toss, sweep the trees with force (of the wind),' with infixed li, may be due to a combination of rasicás shake from side to side,' and a hypothetical *isis.

These formations with infixed la, li, lo have the appearance of the rather numerous class of roots with peculiar final reduplication, the last syllable consisting of the first consonant of the root and the vowel and final consonant of the second syllable, e. g., payid-púd approach gently (of the wind),' probably from the root payid carry something with force (of the wing).' It is not impossible that this final reduplication may be modeled after the pattern set by the forms with infixed la, li, lo.

Examples of roots with reduplicated final syllable and inserted la, li, lo are:

bulitiktik 'be very full,' from the root butiktik be replete with,' with infixed li,

1 salaksák, p. 276a, is evidently not related to this root.

146

Wolfenson, The Infixes la, li, lo in Tagalog.

[1906.

halomigmig something moist,' 'a small mixture; lukewarm',' from the root húmig mix some liquid with another thing,' with infixed lo and reduplicated final syllable. malangotnĝót from manĝotngót a kind of tree,' with the same meaning.

There is a small class of roots in Tagalog composed of the group halo followed by two identical syllables. This halo seems to be a prefix and the two following syllables the ultimate root, e. g., halo-kipkip from kipkip cross one hand or foot across the other.' This prefix halo contains perhaps the infix lo. It is possible, however, that these words are examples of the same formation as halomigmég, in which hámig is the root, or are at least modeled after forms of that character. But in no other case does a root composed of ha and one of the reduplicated syllables, as, e. g., *hakip, occur, to which these forms might be referred.

The reduplicated roots having initial halo are, viz.: halobaybay

small sardine,' perhaps connected with baybáy shore of the sea.'

halobitbit an herb thus named.' A connection with bitbit raise up something in the hand,' is uncertain.

halokipkip fold the arms on the breast,' derived from kipkip 6 cross one hand or foot over the over.'

halotakták tip (metal) of a lance,' derived from takták the iron which caps a staff,' and takták weed-hook; a pole with a piece of iron at the point.'

halotiktik ‘song of the newt, lizard,' probably connected with tiktik 'song of a certain bird.'

In Tagalog then, just as in Bisayan, a number of words containing the infixes la, li, lo are found, the words containing these infixes being treated as roots in the various grammars and dictionaries of Tagalog. In general the particles seem to have little or no force, although in some instances traces of a diminutive meaning are apparent. In conclusion I desire to express my thanks to Dr. F. R. Blake for a number of suggestions and explanations.

This definition is taken from Nigg, A Tagalog English and Eng. Tagalog Dict., Manila, 1904, s.v.

Did the Babylonian Temples have Libraries?-By MORRIS JASTROW, Jr., Professor in the University of Pennsylvania, Philadelphia, Pa.

THE main source for our knowledge of the literature-in the proper sense-produced in ancient Babylonia is still the remarkable collection made chiefly' by king Ashurbanapal of Assyria (668-626 B. C.) which was discovered by Sir Austen Henry Layard in 1849 in the king's palace at Nineveh. Layard came across several rooms in the so-called South-West palace at Kouyunjik (opposite Mosul) filled with clay tablets of varying size. Subsequent excavations and searches for further tablets in the palace in question were made by Rawlinson (1853–55), Rassam (1854, 1877-1883), George Smith (1873, 1874-1876), Budge (1888, 1889, 1891), King and Thompson (1903). Through these combined efforts the number of tablets recovered was considerably increased until at present somewhat over 20,000 tablets and fragments have found their way to the British

1 Bezold in the Introduction to his magnificent Catalogue of the Cuneiform Tablets in the Kouyunjik Collection of the British Museum (5 vols. London, 1889-1899), vol. v, p. xiii, accepts as satisfactory the evidence that the collection existed "in a humble form" in the days of Sargon (722-705), the great-grandfather of Ashurbanapal, and that additions were made to it by Sennacherib (705-681) and Esarhaddon (681-668). See also British Museum Guide to the Babylonian and Assyrian Antiquities (London, 1900), p. 34, and Bezold's suggestive remarks in his article "Bibliotheks- und Schriftwesen im alten Nineve" (Centralblatt für Bibliothekswesen, xxi (1904), p. 273. A tablet like K. 3600—a hymn addressed to Nanâ by Sargon-is almost conclusive evidence in favor of this view. Likewise tablets like KK. 9452 and 9487-copies made by contemporaries of Sargon and Sennacherib.

* Layard, in his account of the discovery (Nineveh and Babylon [London, 1853], pp. 344-347), speaks chiefly of two rooms but mentions also adjoining chambers containing tablets but in far smaller numbers." The Library, it thus appears, was stored in several rooms. The size of the tablets varies (see Bezold, l. c., p. xv) from 15 × 83 inches to 1× inches.

Museum. The size of the collection in connection with the large variety of subjects represented,' together with what we know of the manner in which the collection was made, make it in the full sense of the word a Library, and the designation "Ashurbanapal's Library" has therefore become a general one for this royal collection of tablets-and properly so.

When it became evident that the Library as indicated already in a number of the cases by the subscripts, apart from internal

3

1 of this number about 14,000 constitute the original Kouyunjik collection gathered by Layard, and the rest-marked off by the date or source of acquisition into 29 separate collections-were secured by the subsequent explorations. A small number of tablets from the Library found their way to other museums or into private collections (see Bezold, 1. c., p. xv). Numerous inscriptions and inscribed objects (clay cylinders, clay seals, vase fragments, bricks, obelisks, etc.) and other objects (jars, spearheads, nails, ornaments, etc.) included in the Kouyunjik collection have nothing to do with the Library proper, and there are also quite a number of tablets entered as Kouyunjik inscriptions (as e. g., KK. 6697, 8755, 8860, 8866, 9288, 9599, 9920, 11958; DT. 108, 260; Rm 2. 588, 81-7-27, 205, 209, 210, 213, etc., etc.) which do not appear to have come from that place. See Bezold, "Bibliotheks- und Schriftwesen im alten Nineve (Centralblatt für Bibliothekswesen, xxi. (1904), p. 259, note, and Bezold's catalogue, p. 1952.

2 See, e. g., the survey in Bezold's Introduction, l. c., pp. xviii-xxviii, and Menant's La Bibliothèque du Palais de Ninéve (Paris, 1880), Chap. iii, though this latter work can no longer be recommended as a guide.

3 These subscripts are of two kinds, (1) either a brief indication that the tablet in question is "the property of king Ashurbanapal of Assyria," which, as Bezold (Bibliothekswesen, etc., p. 275) has pointed out, appears to have been stamped upon the tablet and often accompanies (2) a longer colophon furnishing the name and number of the series to which the tablet belongs, a more or less sterotyped form of praise for Ashurbanapal for having followed the promptings of Nebo the god of wisdom and of his consort Tashmitum, (or in some instances other gods, e. g. Shamash and Adad in the case of omen tablets; see Boissier, Documents Assyriens relatifs aux Presages, p. 232, and Cuneiform Texts, etc., Part xx, pl. 33), to gather the wisdom of the ages in his palace, with further references in many cases that a text represents a copy of an older one, or an extract, while in some instances the name of the copyist or owner is added and the source of the original text-Akkad, Babylon, Cuthah, Nippur and Ashur-specifically stated. That more subscripts are not preserved is due of course to the fragmentary condition of most of the tablets; and while it does not follow that all the tablets were provided with more or less explicit subscripts, this was certainly the case with all the tablets belonging to a series and in many other instances. See further Bezold's article (l. c.) p. 275 seq.