rect; as in many other parts of the Mohammedan world it would seem that certain ancient cults and ceremonies have been retained, which the Moslem has been unable to eradicate.

The Gara tribe is divided into families, the chief of which is the Al Kahtan family, and the head of the Al Kahtan family is recognised as the Sheikh by all the Garas. These families have constant blood feuds amongst each other, which they make up when war is on with their neighbours, the Mahri tribe, or the Geneveh tribe. Wali Suleiman during late years has done much to heal these feuds, and it was through his instrumentality that we were enabled to penetrate into the Gara mountains, with an escort of the heads of the chief families, with comparative safety.

(3) The archæological interest in the plain of Dhofar centres chiefly in its connection with the frankincense trade and the towns established in ancient times along the coast by the merchants who provided the ancient world with the odoriferous drugs.

We have several classical authorities who refer to this district, notably Claudius Ptolemy, the author of the 'Periplus of the Red Sea,' Pliny, and a few others. From them we can gather certain definite points, that beyond Ras Fartak and the Sachalites Sinus there stretched a fertile coast line known as the Libaniferous coast. The capital of this district was according to Ptolemy called the oracle of Artemis (Marreior 'Apréjudoç), and the city next in importance was called Abyssapolis, near which was the harbour, the portus nobilis, or Moscha of the Periplus, where the merchants on their way to and from India used to tarry during the violence of the monsoons.

Along the whole line of the plain of Dhofar there are no less than seven spots where ruins occur, all indicating towns of considerable size ; but on close examination of all of them there can be no manner of doubt that at Al Balad and Robat-which are about two miles from one another and connected by a series of ruins-the capital stood. These places are close to the coast, and nearly in the centre of the line of plain, and consist of the remains of many temples, tombs, and public buildings. The acropolis is well marked with the débris of buildings; there is also a tiny little harbour, evidently only for small craft, across which a chain was discovered, the Arabs say, a few years ago. Then there is a moat round the outer edge of this town, in which water is still found, and bulrushes. The columns still standing form an interesting link, which connects these ruins architecturally with the other ruined sites of the Sabæan world; they are square and fluted at each corner, and with step-like capitals. A further development of this is evidently of later origin, when they decorated the capitals with floral and geometric devices. The columns at Axsum in Abyssinia, at Koloë and Adulis on the coast of the Red Sea, and at Mariaba in Yemen, are all of the same character, and indubitably establish the Sabæan origin of the ruins. One column at Robat we found with a capital decorated on four sides, three sides with intricate geometric patterns, and the fourth with the Sabæan letters ein and 7' alternately. No other ruins either in size or architecture on the plains of Dhofar can compare with these, and we can safely say that they formed the capital of the district, which Claudius Ptolemy calls the oracle of Artemis (Μαντείον 'Αρτέμιδος), and which in later times was known as Mansura, where dwelt, Yakout tells us, the Prince of Dhofar, who had a monopoly over the frankincense trade, and punished the infringement of it with death. In later times the Persians occupied this spot, in the fourteenth and fifteenth centuries of our era. Το them we owe the fact of the disturbance of the old Sabæan columns, and

the utilising of them to erect mosques, many of which are now standing in a fair state of preservation.

We tried to find the site of the oracle of which Ptolemy speaks, but could not come to a satisfactory spot until we visited the mountains, and in the Wadi Nehaz, about 9 miles from the capital, just at the foot of the mountains, we found a curious natural hole, about 150 feet deep and 50 feet in diameter. Around this there was a wall of Sabæan origin which had a massive gatepost, and in the immediate vicinity were traces of many ruins. For several reasons I am inclined to believe that this is the site of the oracle mentioned by Ptolemy. In the first place, the hole resembles in character the site chosen for an oracle in the ancient world, bearing a remarkable resemblance to the holes which existed in Cilicia, the oracles of the Corycian and Olbian Zeus, and several other spots in Greece. Secondly, Yakout tells us that the abode of the Adites was half a day's journey from Mansura, the term Adites generally being given to the adherents of the ancient cult; and, thirdly, because there is no other spot on the plain of Dhofar where one can say there is a probability of an oracle existing.

Yakout further tells us that 20 parasangs from Mansura was the excellent harbour, frequented by the crafts on the way to and from India, and by merchants in search of frankincense. The author of the Periplus refers to this harbour, and calls it Moscha, and Ibn Khaldun also speaks of it as Merbat. As we journeyed along the coast we were constantly on the look out for this harbour, and on the second day, after leaving the ruins of the capital, we reached the village of Takha, in the vicinity of which are traces of many ruins scattered about, but inferior in architecture to those at Al Balad. Next morning we were conducted by the natives round a headland, and there saw a long sheet of water stretching inland, but silted up at the mouth by a sand belt, over which the sea flows at high tide. This same sand belt now separates from the shore a rocky island with traces of fortifications on it. There can be no doubt but that this is the harbour, and the island rock guarded the double entrance to it before the invasion of the sand. The harbour is deep, and extends inland about a mile and a half, and there are many ruins around it. Here we have the portus nobilis of the Periplus, the harbour to which the frankincense merchants came, and it is, as Yakout tells us, just 20 parasangs from the capital. The term Moscha, given to it in the Periplus, is a common term given to bays and inlets on the Arabian coast. Merbat, the name given it by Arabian writers, is still retained in the headland 12 miles east, where Arab dhows find a shelter during the north-east monsoons, but it affords no other harbourage, and Ptolemy's name for this place is Abyssapolis, a name which I consider to be derived from the great abyss which I have already described as existing a few miles inland, and which must have been a conspicuous and well-known object to all merchants who frequented this port. Ptolemy, as it will be seen from the name given to the capital, gave Greek names, or equivalents, to the places on this coast, and in naming this place he evidently used the most conspicuous object in its vicinity. Thus we were able to reconstruct on fairly probable lines the geographical features of this frankincense district, and fix the position of the sites of its towns.

Calibration of Instruments used in Engineering Laboratories.—Report of a Committee, consisting of Professor A. B. W. KENNEDY, F.R.S. (Chairman), Professor J. A. EWING, F.R.S., Professor D. S. CAPPER, Professor T. H. BEARE, and Professor W. C. UNWIN, F.R.S. (Secretary). (Drawn up by the Secretary.)

Ar the first meeting of the Committee it was decided to investigate initially the accuracy of instruments for measuring the tension coefficient of elasticity, or Young's Modulus. A general letter was addressed to various professors and others in charge of engineering laboratories inviting co-operation. Most of those written to agreed to make a series of measurements for discussion by the Committee.

It was then decided that sets of standard test bars should be prepared, to be subjected to tension and measurement. Figs. 1, 2 show the forms of test bar decided upon. Two of the standard bars of each set are cylindrical bars, with screwed ends of about 1-inch and 3-inch diameter. These have gauge points for measuring instruments, suitable for extensometers of 8-inch, 10-inch, 16-inch, or 20-inch range. These bars are of a special steel of high tenacity, rolled specially for the Committee by the Blaenavon Company. The whole of the bars were cut from a single rolled bar about 20 feet in length, and were very accurately turned to the required dimensions by Mr. W. R. Munro. The third bar of each set was a flat bar, of section about 2 inches by inch, of mild steel. All these bars were cut from a single plate, and they were prepared with gauge points at 8 inches and 10 inches.

In order to obtain some preliminary information as to the mechanical properties of the standard bars, one round bar and one flat bar were tested in the testing machine at the Central Technical College. following table gives the results obtained :

Preliminary Tests of Materials used for Standard Bars.

TENSION EXPERIMENTS.

The

The Committee then drew up a test-sheet form to be issued with the bars, on which measurements were to be recorded. These sheets were so arranged that two sets of measurements for each bar, and the mean of these, should be recorded; also that the extensions for short and long ranges of stress should be recorded. It was hoped that in this way some measure of instrumental errors would be obtained.

In January two sets of bars were sent out for measurement, to be circulated amongst those who had consented to co-operate with the

1005

KK

Committee. The measurements have taken much time, and the whole of the reports of observers have not yet been received. Some of those already sent reached the Committee too late for discussion this year. It appears, therefore, to the Committee that it is unavoidable that only an interim report can be presented this year. The Committee ask for reappointment, in order that the results obtained may be discussed and presented.

It may, however, be useful even at this stage, to give a very short summary of some of the results sent in. The following table gives a summary of the measurements of the standard bars by different observers. It will be seen that there is an appreciable difference even in the measurement of the dimensions of the same bar. The flat bar, it may be noted, was machined on the edges only.

The following table gives the values of the coefficient of elasticity for the greatest range of stress observed for each bar. The values are given for the first and second loading of the bar, and also the mean of the two observations. It will be seen that, even for the mean of two sets of readings, over the greatest range of stress the elasticity of the bar permits, there are very appreciable differences in the values obtained. It remains to be considered in a more detailed discussion of the results whether any evidence can be found as to the source of the discrepancies. It may be due to error of the testing machine, to error of magnification by the extensometer, or to error of calibration of the extensometer. It is just possible it may be due in part to temperature or other action independent of both testing machine and extensometer.