more than this as a vocabulary should be in regular use by them. The teacher can do the rest-it pertains to translation (see on subdivision (1) ); or the pupils may profit by occasional references to a large Latin Dictionary.

(b) It has been a practice on the part of some makers of text-books to give, without discrimination, references to explain some rather unusual construction and others which merely name the most ordinary usages. When a boy fumbles the pages of his Grammar to seek out a certain number and finds as a result only "The ablative is used to denote the means or instrument of an action', or 'The dative is used as the indirect object of verbs', he is not likely to repeat the process many times; he will endeavor not to get 'stung' again. The result is that at other places he will fail to get needed information about constructions which, perhaps, he has never met before. If, on the other hand, the editor would always explain constructions in his note, and give references to the Grammar only for those which are unusual, then the pupil could judge for himself whether he understands it well enough without fingering the Grammar. An objection will be raised to this, that the editor often wishes to put it in the power of a student to find out how to construe a certain word, without telling him outright before he has a chance to think. But is any pedagogical good gained by this cryptic method at all comparable with the harm that is done? Is not the same end attained if the editor says, for example, 'tactis: construe with such and such a word', or 'What use of the ablative is this?', thus saving the Grammar ammunition until the eyes of a real enemy are seen? If the editor feels that a 'completely parsed' text is needed, in competition with those already on the market, he might put it in a separate volume; but probably living teachers will prefer to keep it out of the hands of their pupils.

In summary, the following principles have been enunciated. The utmost initiation into the thought and the real conditions of life of the period of antiquity should be included. Extraneous matters should be excluded, lest attention be spread too thin. Nothing should be included which will not be used, in order that habits of neglect and inattention may not be fostered. Symbolism should be reduced to a minimum, both in the definition of words and in the naming of construction, in order that opportunity may be left for the products of real thought. HASTINGS-UPON-HUDSON.

BARCLAY W. BRADLEY.

this time2. In view, however, of the general interest they have aroused and of the somewhat uncertain nature of the reports which have appeared from time to time in American, as well as foreign, reviews, it may be that this brief notice of the more important of his discoveries since he began work on the Flavian palace in October, 1911, will have a temporary value for readers of THE CLASSICAL WEEKLY.

Investigations have been limited strictly to the great series of public rooms which occupy the center of the hill-top, formerly erroneously known as the domus Augustana, but nowadays, from its topmost archaeological layer, commonly referred to as the domus Flavia or palace of Domitian. I shall call it, for the sake of convenience, the imperial palace, or simply the palace, though this term may of course be legitimately extended to cover most of the hill-top. The discoveries, in any case, may be conveniently grouped under three heads: first, substructures, pavements, and accessories of the palace in the successive stages of its development; second, remains of republican houses over which the palace was built; and third, remains of a primitive archaeological stratum beneath the palace foundations.

The palace proper, as Boni has shown, rests on the summit of the primitive Palatium, not on the intermontium between Palatium and Cermalus. He has succeeded also in exposing the substructures of the tablinum in such a way as to present a fairly definite chronology of its development. The cement foundations of the domus Flavia or palace of Domitian, containing silex, cut older ones containing travertine, probably to be assigned to the palace of Nero. These in turn cut still older walls containing fragments of republican tiles, probably of the palace of Caligula. This can be seen to advantage from a point near the entrance to the tablinum.

To the palace of Domitian Boni attributes the magnificent granite pavement, with its border of Numidian marble (giallo antico), in the triclinium. The huge 60-foot octagonal impluvium of the atrium, or peristyle, cut through but not laid bare by previous excavators, is likewise a part of the Flavian palace. To the Neronian structure, however, belongs a series of five apartments in opus signinum under the basilica, which, on the evidence of the action of salt on the plaster, Boni identifies as a reservoir for salt-water fish destined for the imperial table.

At a depth of some four or five meters under the triclinium of the Flavian palace a series of rooms has been exposed, which, Boni thinks, forms a part of the house of the emperor Tiberius. This, in his opinion, constituted the kernel of the later palace, which was built over it at a much higher level, seemingly in order

The preparation even of this modest report would have been impossible without Commendatore Boni's courteous aid. See, moreover, his own recently published informal statement, Les nouvelles découvertes du Palatin, Brussels, 1914 (a lecture delivered at the University of Brussels, June 18, 1913), and E. Steinmann, Sul Palatino, in Nuova Antologia for March 1, 1914, pp. 133-140.

to make it even with the higher grade of the hill-top immediately under the atrium. There are remains of pavements with floral designs in colored marbles, besides a structure which suggests to the excavator something in the nature of a water-organ.

The much talked of elevators, which Boni believes carried passengers between the summit of the Palatine and the Circus Maximus, consist of three, or possibly four, rectangular pits (not twelve, as published in certain earlier reports, e. g. American Journal of Archaeology for 1913, p. 117, probably in confusion with those of the Forum), situated in different parts of the hill. One is in the northern nymphaeum of the palace, one under the Villa Mills, one in the so-called domus Liviae, with perhaps a fourth under the socalled stadium. The only one which has been even partially excavated is that in the nymphaeum. It is about one and one-half meters square with its sides parallel and at right angles to the lateral walls of the Circus. It has been emptied to a depth of 36 meters, i. e. within about six meters of the Circus level, without reaching bottom. Near it, at a depth of ten meters, was found a chamber containing the remains of a stone base, which Boni thinks supported an hydraulic engine providing power for the mechanism. The end of a bronze cylinder there discovered, and a water reservoir above, constitute the evidence thus far adduced. Dating is as yet out of the question.

Of the republican houses discovered under the palace, the most important from an historical standpoint is the one near the northern nymphaeum, which Boni has identified by the inscription on a marble cornice, as yet unpublished, as the dwelling of Tiberius Claudius Nero, the husband of Livia. The evidence for this identification will of course be awaited with keen interest by those who have followed the mooted question of the so-called domus Liviae. In any case, the house contains splendid fragments of pavement in floral designs of red and green porphyry with giallo antico, besides interesting mural paintings on Homeric subjects the dispute between Agamemnon and Achilles, and the quarrel of Ajax and Ulysses over the

arms.

Remnants of two houses have been found in the immediate vicinity of the Neronian fish-reservoir under the basilica. The later of the two, dating from the last years of the republic, is cut by the walls of the reservoir. It, too, is important for its frescoes, which show symbols and ornaments imported from Egypt and Asia Minor. The precious blue from Alexandria is used, as well as vermilion, perhaps from Monte Amiata in Tuscany. To this house belongs the painting of Aegisthus, Clytemnestra, Electra, and Orestes, discovered in the Farnese excavations early in the eighteenth century. Of the older house only a stairway in opus incertum remains.

The fourth and best preserved of these republican houses is that which lies under the lararium and the adjoining portion of the tablinum of the domus Flavia.

The upper story, which, on account of the slope of the hill, contained the atrium with its still existing impluvium, was originally paved with local calcareous stones and flints in various colors. When the house was rebuilt, however, at the time of Augustus, with a somewhat different orientation, this mosaic flooring was covered over with a new pavement of African breccia, Euboean cipollino, and a gray and white Greek marble. The lower story of basement rooms, however, to which one may now descend by the original staircase of seventeen steps in tufa blocks, seems to have been given up at that time, for it was found filled with architectural fragments and rubbish from the upper portion of the republican house. The débris cleared away, there remained an imposing suite of rooms with frescoes and marble pavements in geometrical designs of colored cubes, and handsome griffins in stucco relief. The house was manifestly a rich one, and Boni's suggestion, based on purely topographical grounds, of an identification with the domus Catillinae, will doubtless serve to give it a popular name.

Under this house were found terra-cotta decorations of a still older dwelling of the third century B. C., and, in a series of galleries cut in th. tufa rock, early vase fragments of the sixth and fifth centuries, some indigenous, some of the white Campanian type, some of Etruscan black bucchero, and some of the Attic redfigured style. Graffiti on these fragments are in early Latin, Greek, and Etruscan, and as a whole they serve to throw light on Rome's relation with the outside world before the Gallic invasion of 390 B. C.

Interest in the primitive remains which have been brought to light under the substructures of the palac centers first of all in what Boni calls the Mundus, the famous sacred pit which on three specified days in the year, August 24, October 5, and November 8, stood open to afford egress to the denizens of the lower world. And indeed, in the northern portion of the atrium, near certain indications of primitive hut-foundations sunk in the tufa rock, was unearthed the half of a square slab of tufa which originally contained a circular hole in the center, the very stone, Boni thinks, which gave the name of Roma Quadrata to the place (compare Festus, 258 ff.: Quadrata Roma in Palatio ante templum Apollinis dicitur, ubi reposita sunt quae solent boni ominis gratia in urbe condenda adhiberi, quia saxo munitus, est initio in speciem quadratam). Under this slab, moreover, was discovered a cylindrical chamber with a cone-shaped roof, a sort of tholos, which, by its shape, has been identified by the excavator as the Mundus proper (Cato ap. Festum, 157 ff.: Mundo nomen impositum est ab eo mundo, qui supra nos est. Forma enim eius est, ut ex his qui intravere cognoscere potui, adsimilis illae). And in the bottom of this a wide circular pit leads downward to a series of subterranean passages cut in the rock twelve meters below the surface, in which we are invited to recognize the inferiorem partem consecratam Dis Manibus of Festus 157.

These catacomb-like galleries are as yet only partially excavated, and have so far yielded nothing beyond two ordinary republican vases, a few bronze bases, and part of a bronze helmet ribbed with iron. The explored galleries are some three meters in height and one in breadth, and the walls are smoothly polished. They converge at one point into a circular cavern from which a second circular pit leads to the surface just west of the wall which separates atrium and tablinum. Boni compares them with the favissae of the Capitol and with similar underground corridors found along the Latin coast near Nettuno. Their polished walls and complicated plan preclude the idea, he thinks, that they are early quarries, and he regards them as sacred granaries of the primitive state. It must be added, however, that other pits lead to other favissae at higher levels, notably under the various republican houses. Boni thinks in these cases of domestic larders. The whole matter is of course of vital interest and will attract wide and serious attention when the evidence is published in full.

Meanwhile, under the Commendatore's genial direction, the Orti Farnesiani have been set out with a magnificent display of classical flora which promises ere long to be in bloom. A glance at the plan accompanying Steinmann's article, alluded to above in note 2, shows among other species the laurus nobilis, myrtus Romana, buxus sempervirens, hedera helix, cytisus, laburnum, acanthus mollis, hyacinthus, narcissus, thymus, verbena, and arbutus. Thus even the unarchaeological Roman en promenade must needs add his meed of thanks to the archeologo-giardiniere', whose gardens already furnish a pleasant contrast to the temporary roofings of canvas and tin which shelter the new diggings.

REVIEWS Dominance et résistance dans la phonétique latine. Par C. Juret. Studien zur lateinischen Sprachwissenschaft I. Herausgegeben von M. Niedermann und J. Vendryes. Heidelberg: Carl Winter's Universitätsbuchhandlung (1913). Pp. xii +263. 7 Marks.

This extensive monograph has as its aim the determination of the effect of the position within the syllable upon the development of the original sounds into classical Latin, and, in this connection, the bringing into order of the rules of development of the sounds. In other words, the author's problem is this, if we may take a single example as typical: does original t have the same development, when initial in the syllable, whether that syllable begins the word or begins in the middle of the word? Does it have the same development, at the end of the syllable, whether that syllable ends in the middle or at the end of the word?

A thoroughgoing examination of M. Juret's work would involve the recapitulation of an immense amount

of technical detail, and the reviewer feels obliged, so far as possible, to keep to the discussion of general principles-save for sample minutiae.

(1) M. Juret finds (19-95) that the consonants reduce themselves to the following rule: In like positions in the syllable, regardless of the position of the syllable in the word, a given original consonant develops in the same way-except that original aspirates develop into voiceless sounds initially in the word, and into voiced sounds at the beginning of non-initial syllables, with some differences when intervocalic. Some special groups, like medial -ld-, show special development.

Such a simplification of the rules is very gratifying; in the main, M. Juret does not run counter to accepted phonetic laws. We may agree with him (66-68) in rejecting the old formulae that initial tw- and qwbecome p- and - respectively; detailed refutation of these laws had been given shortly before by Persson, Beiträge zur Indogermanischen Wortforschung (Skrifter utgifna af K. Humanistiska VetenskapsSamfundet i Uppsala, x, 1912), pp. 470-479 and 520535. But we cannot follow him in rejecting (45-47) the formula that dr becomes tr, when we have such examples as taetro-, uter utris, citrus; nor in rejecting the rule (71) that in intervocalic -sw- the s is lost with compensatory lengthening of the preceding vowel, and, if uv result, the v is lost and the vowel shortened; for his explanation of pruina fails of cogency when he assumes that it arises by dissimilative loss of the second r in *prurīna-it is normally the prior and not the posterior consonant which disappears in this way, and the verb pruriō shows the combination remaining intact. In casting aside the formula (79-81) that original final -t becomes -d in Italic, M. Juret fails to heed properly the testimony of Oscan, where the IndoEuropean secondary ending for the third person singular active in verbs is regularly -d, while the primary -ti has been protected from the change by the vowel, which was later lost, and appears in Oscan as -t.

(2) Juret finds (96-104) a similar status of affairs in the vowels of medial and final syllables. However, one hesitates to follow him in considering the change of final -et to it as merely analogical (97), though the -eof tutimet must be reckoned with. In discussing (101) laudat and meritod, where the long vowel shortens before final -t, but remains long before d-, which is lost, he seems to raise difficulties which do not exist. treatment of final -à -5 -u (103-104) is quite unsatisfactory, in view of the Oscan and the Umbrian forms of the neuter plural, sequere and ipse, and cornu.

His

(3) In medial syllables, M. Juret entirely rejects (116-132) syncope of a short vowel after a stop or s; his explanations of those words ordinarily supposed to have syncope in this position lead him, in his own words (131), to what "paraît une vraie débauche d'étymologie". In particular, others will be unable to accept his interpretation of some or all of the following words: quindecim, hospes (117); sumo, mixtus, dexter (118); supra, extra, intra (121); cette (131);

sumo, pono (151-152). His conclusions seem invalidated.

(4) M. Juret next (133-153) sets up the following rule: A short vowel is absorbed (= lost) only after a doubled sonant consonant (= r l m n v) or after a sonant consonant preceded by a long vowel. Again, we can hardly accept the limitations of his theory; balneum (135), palma (135), indulgeo (136) seem to resist his explanations.

(5) Relative to 'syncope' and samprasārana', M. Juret sets up (153-170) new rules which would require too much space for full citation; he entirely rejects current views. One typical instance may be cited: quantillus he derives from *quant(o)-lolo-s, with 'suppression' of the first o for reasons of rhythm; vowel weakening produced *quantlilus, whence quantillus came by metathesis. Comment seems unnecessary.

(6) In the consideration (171-191) of the loss or retention of the short vowels of final syllables, however, he makes it worth considering whether the loss of -iin the nominative singular of i-stems, as in mors, mens, may not be analogical merely to lux, dens, etc., with their dissyllabic genitives.

(7) Lastly (192–261), he disputes the development of v intervocalic, particularly against Solmsen, whose rules for the loss of the sound he attempts to controvert (193 ff., 251 f.). He develops these rules: In the group formed by vowel ++ short vowel + consonant of non-final syllable, v is lost and the short vowel is contracted with the preceding vowel, if the short vowel is followed by n, t, s, r, velar or palatal I, y. When other sounds follow the short vowel, the v remains and the short vowel is absorbed.

Similarly, he sets up the rules: In initial syllables, ve becomes vo (and v is lost if standing after a consonant) provided n, r, t, s, h, or palatal g begins the next syllable; ov in initial syllables becomes av, except before i; ev in initial syllables becomes ov, except before i.

But other explanations are readily discoverable to remove the difficulties which he finds under the accepted rules; and he is not always correct in his statement of facts: doublets are not all too infrequent in Latin (despite 113); rio is only typical of a whole class of words in Italian (cf. 245); the rate of speech, Sprachtempo, is a factor which must not be left out of consideration (cf. 112-113). Other problems are left unsolved: why do we have clavaca, as well as clovaca and cluaca? and did novus novem keep the original e until after *covos had become cavus?

Misprints are regrettably frequent, two occurring even in the list of errata on page xi. Certain errors of fact occur. Tibicen (sic!) is given at 119, 10 as an instance of a word with medial, though the i's are both long. Quattuor does not come from *-twōr (65, 12); cf. Brugmann, Grundriss der vergleichenden Grammatik der Indogermanischen Sprachen, II, 22, 13. On 120, M. Juret says that anaptyxis occurs only

before velar 1, and forgets the whole class of words with palatal, of which stabilis from *stablis is an example. His analogical proportion on page 125 will not give reperio as the fourth member. Venentficus, cited on page 147, is not actually found. When on page 191 he denies an i-stem to vigil, because of vigilare instead of *vigiliare, he forgets piscis and piscari. On page 33, he derives farnus from *farznos < *farxnos, which despite fraxinus cannot be correct, for rksn > rsn> sn > n, with compensatory lengthening of the preceding vowel; cf. mantele < *-terk-sli, from the root seen in tergère, and cena < *kertsna: farnus must be from *fargnos or *farginos, or possibly from *farksinos. On page 57, his explanation of refert cannot be maintained, since Persson, on pages 294-305, of the work referred to in the fourth paragraph of this paper shows that Walde's law of loss of the aspiration of an initial aspirated stop before a liquid when an aspirate ends the root or begins the next syllable is without foundation, and that the loss is due merely to the fact that a consonant immediately follows.

It is not surprising if so radical a revision of the phonetic laws of a language already so well studied as Latin should at this day fail to carry conviction. The main lines of development are too clear for anyone to hope to break new paths except in details. But, apart from the warning which M. Juret's book serves to give, we find in it many keen observations on individual words-some 200 of which are listed in an index because of original suggestions or special comments; on some of these he will doubtless win acceptance, though the student must constantly check up his views by reference to Walde's Lateinisches Etymologisches Wörterbuch: we may call attention especially to what M. Juret has to say on cunctus, acerbus, saltus, sinister, iuxta, doctus, tostus, vir, pulmo, dulcis.

CHEMUNG COUNTY CLASSICAL
ASSOCIATION

Following out the plan of the Classical Section of the New York State Teachers' Association a meeting of the classical teachers of Chemung County was held at Elmira College on May 2. The program was as follows: Remarks on the Purpose of the Conference, by Professor H. A. Hamilton, of Elmira College; The Socialization of the Classics, by Dr. Mason D. Gray, of the East High School, Rochester; The Teaching of Latin to Younger Pupils. by Miss Laura C. Manley, Elmira Free Academy; Round Table on the Problems of the Latin Teacher, led by Principal F. R. Parker, Elmira Free Academy.

ELMIRA COLLEGE.

H. A. HAMILTON.

A Roman School, Miss Susan Paxson's Latin play (see THE CLASSICAL WEEKLY 5.1) was given on May 1 by Latin students of the Oakwood Seminary under the

The forty-sixth annual meeting of The American Philogical Association will be held at Haverford College, Haverford, Pa., on Tuesday, Wednesday, and Thursday, December 29-31. The Executive Committee, to which the place of this year's meeting had been referred, with power, carefully considered two invitations-one to meet, with the Modern Language Association at Columbia University, the other to meet with the Archaeological Institute of America at Haverford. In view of the feeling of many members that the Association should meet with the Institute, provided the business of the Institute can be kept out of the way of the reading of papers, the Committee decided to accept the invitation from Haverford.

The first session of the Association will be called to order on Tuesday, December 29, at 3 o'clock. In the evening, at the joint session of the Association and the Institute, Professor Edward Capps, President of the Association, will deliver the annual Presidential address.

Further details will be communicated in a circular soon to be issued (or already issued) by the Secretary of the Association to all members. Others interested may secure copies of this circular by applying to Professor Frank Gardner Moore, Columbia University.

Mr. James Loeb, formerly of Kuhn, Loeb and Co., New York, writes as follows:

The great and legitimate aim of a business man is to make money, to provide for himself and his family such luxuries and comforts as his tastes and social standing demand. But when a man has reached the goal of his desires, when he has 'made his pile' and wants to enjoy it, then comes the time for the making of the real and only Balance Sheet. Then he must ask himself, 'What are my resources, now that I have everything that money can buy? What are my spiritual and intellectual assets? How can I best spend what is left to me of life?' Lucky is the man whose early training fits him for something more than the golf-field, or the tennis-court, or for something better than the gaming-table when his days of business activity are over. He can taste the gentler pleasures that await him in his study and by the blazing hearthfire. His Sophocles or his Homer or his Catullus will make the winter of life seem like its early spring when the greatest struggle he knew was with the elusive rules of grammar and syntax.