Beneath the kid is the most interesting figure of the composition. It is a human shape with uplifted hands, the lower part of the body terminating in the tail of a fish (?), and on the back a projection looking like wings.

Behind the sacrificer and facing the same direction, is another deity with uplifted hands, presumably the wife of the Sun god. Between this deity and the sacrificer are two other figures: above, a beast lying down; below, a composite creature, fish and man, holding the hands aloft.

The seal 3 is clearly the same, though only a small part is to be seen, since the impression is made across the written surface. But the winged figure with the kid above it removes all doubt.

Regarding 34, 3e the case was at first not so clear, and I long thought that these came from a fourth seal. The central figure in 3d with the man-fish before it, and with the hair like a curved cue behind the back, is the same as on 3. Moreover, there stands before this figure a priest offering a kid, as in 3, 3a, though only the legs of the kid are preserved, the body being lost by a small break. The winged figure is similarly lost. The seated deity on the right is not given in this impression, but in compensation we have four figures on the opposite side of 3a, namely a standing deity turned toward the left; near his shoulder, a head, that of a demon perhaps; below, the long body of a serpent; and below this a kneeling, diminutive figure in human shape, looking toward the spectator. head of the sacrificer are to be dimly seen the fore-legs of the kneeling animal, clearly seen in 3 and 3a. There is no doubt therefore that 3d is the same as 3.

Just behind the

Of 3 and 3o little is discernible, but quite enough to remove all uncertainty as to their identity. Of 3 note the lower part of the kneeling figure, and of the two standing figures before him. Likewise the serpent above the kneeling figure. Of 3 may be seen the serpent, the man-fish, and the uplifted hands of the goddess who stands between them.

Of the seal impressions on side F, three (4, 6, 5), being made on the written surface, are too faint for description. No. 4 seems to contain two deities, and between them a zigzag symbol of the lightning. The identity of 6a with 6 seems almost certain, in spite of the faintness of the impression; 5a seems also to be the same as 5. Fortunately 5, 6, and 7 are clear.

No. 5 is a sacrificial scene, in general similar to 3, though the god in this case is standing. In his extended right hand is a rod, and his right foot is slightly raised, as if resting on some object. Above the kid which the worshiper presents are the sun and the moon, and behind the worshiper a goddess with hands erect. Behind her, a crude tree and part of another figure.

No. 6 contains an inscription in four lines, on the right of it Adad, the weather god, and on the left Shala his spouse, each with back to the inscription, and therefore standing face to face, as was the case in seal 1. (See 1a, 1′.)

I have left till the last the most complicated seal of all (No. 7). The impression measures about one inch high by about 14 inches wide. The seal is divided into two registers, an upper and a lower, separated by that twisted rope-like ornamentation which is occasionally found on seals, and which often enters into the decoration of larger works of art.

The lower half shows traces of an inscription in at least two lines. There are two Ea-bani figures contending, and between them a fish. There is likewise a well shaped nude figure of a man contending with an animal.

In the upper half the central figure is a god facing the right, in his uplifted right hand a rod, as if in the act of striking, or more likely a lance, since it is held not by the end but by the middle. In front of him is a small animal running towards him. Then a door in which stands a human shape with hands aloft. Then a fox running toward the door, and on the extreme right a human figure. Below the door, an obscure shape, an animal (?) possibly, also an obscure figure below the fox, and another near the face of the god with uplifted lance.

Behind this god are four other figures, an ox lying on the ground, a diminutive animal above the ox, then the legs and body of a large bird, and lastly a mouse or a rat of unmistakable form.

I have gone thus into detail because it seemed worth while to illustrate by one example the old Babylonian method of sealing tablets. The significance of a seal impression was the same as among us today. It was an additional security. Its importance among the Babylonians was even greater than among us, because, owing to the absence of curves in the writing, the chirography of a man was not so characteristic. Hence the necessity of a large number of witnesses to an important trans

action, sometimes a dozen or more. As a safeguard against forgery, witnesses as well as contracting parties of both sexes affixed their seal impressions to the record. Names of men might be forged, but not impressions of their seals.

It may well be that seal impressions not only had this business value of identification and legitimation, but that they were also thought of as possessing talismanic virtue, particularly when, as was usually the case, they contained representations of the gods.

But why duplicate a seal so often, and why cover all blank spaces and even written spaces of the tablet with seal impressions, sometimes to the extent of making the record in places illegible? Perhaps as an additional safeguard against tampering in any way with the record.

But of greater interest than these questions are the seal impressions themselves. In a pictorial way they tell us much regarding the costumes, natural history, religious beliefs and practices, and the development of art at the time when the seals were cut; and we shall not go far astray, I think, in assuming that a seal is not much older than a tablet on which its impression occurs. The study and publication of such seal impressions has therefore an importance hardly inferior to that of the transactions recorded on the tablets.

In one respect dated seal impressions are of fundamental importance, in that they furnish the only sure means for classifying the seals themselves, which exist in such large numbers in the great collections. Seals of different dates will reflect the changing ideas and tastes of the times, and those from different parts of the country may show enough difference of character to justify us in speaking of different local schools of art. But for many of the seals brought from the East, it is impossible to learn their provenance, owing to the conditions under which they are found and exported. A comprehensive study of seal impressions on dated tablets of known provenance is therefore most desirable, since it may furnish data sufficient for both geographical and chronological classification of the seals.

It is believed that the large collections of tablets in the museums would furnish adequate material for such study. One of the most valuable contributions to the growth of Assyriology would be a volume giving in photographic reproduction a large selection from the seal impressions on dated cuneiform tablets.

The Infixes la, li, lo in Tagalog.-By Louis B. WOLFENSON, Johns Hopkins University, Baltimore, Md.

In the Philippine languages most derivative words are formed by means of the various kinds of reduplication,' and by combination with derivative particles. The derivative particles are very numerous—so numerous in fact that these languages are sometimes spoken of as "particular" languages. They are for the most part prefixes, but there are some few suffixes and infixes. In Tagalog the only recognized infixes are um and in, um forming active verbs, e. g., from the root súlat 'write,' s-umúlat to write,' and in forming the preterite and present of passive verbs, e. g., s-in-úlat 'written,' from the same root.

In Bisayan, in addition to the verbal particles um and in, employed as in Tagalog, we find the infixes la, li, lo inserted after the first syllable of words or roots, the vowel of the particle being in many cases the same as that in the first syllable of the root, e. g., salawayon that which merits correction, reprehension,' from the root saray correct, reprehend;' tilimanan 'mark, sign, signal' from timaon, meaning the same; and tolonan a lesson to be learned, read or memorized,' from the root toon teach and learn, instruct.'

6

6

These particles are used in the following manner. In the Cebuan dialect they seem to form adjectives and frequentatives, as indicated by the statements in the grammar and dictionary. In the grammar of the Cebuan dialect written by the Recolet Friar Zueco, the statement is made' that adjectives of quality, signifying an accidental property of a person or thing to which they are applied, are formed by inserting these particles, e. g.,

3

1 For the various forms of reduplication in Tagalog see W. G. Seiple, Polysyllabic Roots with Initial P in Tagalog, JAOS. vol. xxv, 1904, pp. 287. 288.

* The Cebuan dialect is one of the four principal Bisayan dialects, viz., Cebuan, Hiliguayna, Harayan, and Samaro-Leytean. Cf. Dr. Frank R. Blake's paper, The Bisayan Dialects, JAOS. vol. xxvi, 1905, p. 120 ff., especially p. 123 below.

& Método del Dr. Ollendorff . adaptado al bisaya, Manila, 1871. Op. cit., p. 146 below.

6

talamayon despicable,' silingbahon venerable; adorable,' and tulumanon 'that which is practicable.' In Encarnacion's Bisayan Dictionary,' li and lo are defined as particles which are infixed after the first syllable to form frequentatives. A frequentative meaning, however, is often not very apparent, the derivative with infixed particle having at times a meaning very close to or practically the same as that of the word from which it is derived, as in the case of tilimanon mark, sign,' which has the same meaning as the underived word timaon.

In partially reduplicated roots with these particles infixed after the first syllable, a diminutive force is clearly evident, not only in the Cebuan dialect, e. g., in tolotigolong living being somewhat old, approaching old age, oldish,' from tigolong 'old of both sexes,' but also in the dialect of Samar and Leyte, the Samaro-Leytean, in which these particles appear as ra, ri, ro," e. g., karokabayo 'little horse,' from kabayo 'horse,' and barobalay 'little house,' from balay 'house.'

The diminutive force in these cases may be due to the reduplication of the root; for in both Tagalog and Bisayan there are instances of diminutive reproduction, e. g., Tagalog mabutibúti pretty good,' and Bisayan tawo-tawo little man.' But it is perhaps more likely that the diminutive force is due to the combination of the partial reduplication and the infixed particle.

The infixes la, li, lo are found not only in Bisayan, but occur also in Tagalog in a considerable number of words, although they are not recognized as such by the Spanish grammarians, e. g., salaysay explain' from the root saysay explain, clear up,' 'with infixed la; bulitiktik be very full,' from butiktik be replete with,' with infixed la; and dulotdót dig a little,' which is probably derived from dotdót stir with the finger,' with infixed lo.

6

The force of these particles in Tagalog is for the most part very indistinct, the meaning of the derivative with infixed particle being often the same as that of the word or root from which it is formed, as in the case of the first two examples cited

1 Diccionario bisaya-español y Diccionario español-bisaya, 3a ed. Manila, 1885, pp. 199, 211".

On the correspondence of in the Cebuan and Hiliguayna dialects with r in the Samaro-Leytean, see Dr. Blake's paper, cited above, p. 125. 3 The definitions of Tagalog words in this article are based on Noceda, Vocabulario de la Lingua Tagala, Reimpreso en Manila, 1860.

For the a in the first syllable cf. p. 145.