Zab. The name of this city is written in cuneiform →×N ➡ ➡Y (alu) Arba ilu, (city) Number Four God'. That is to say, the numeral Four is here employed as a divine name or title. This is beyond question. Would it not, then, apart from further evidence, be at least very probable that in Ķiriath Arba', City of Number Four', the numeral is in like manner used as the title of a God, even though the specific symbol of Deity is not employed in the Hebrew as in the Assyrian? But further evidence, is abundantly forthcoming. One thinks at once of the Beth Arbel ( a) of Hosea x 14, perhaps situated near Here we have the name Arba'-ilu taken Assyrian or Babylonian, since the y of the Turning to Prof. Winckler's Geschichte,

Pella on the east of Jordan. apparently directly from the Hebrew Va is wanting.

I find that he adopts this explanation of Kiriath Arba', and further explains Beer Sheba' in like manner as 'Well of Number Seven God'. Thus new light is thrown upon the subject. A God Sibitti, i. e. 'Number Seven', was known to the Babylonians at the period of the First Dynasty. Thus, for example, we find such names as

Arad- (ilu) Sibittim, servant of (God) Sibitti '.1

1

It can

The meaning of Four and Seven as divine titles is elucidated by the well-known fact that the name of Sin, the Moon-god, is commonly written in cuneiform ««<, i. e. "God Number Thirty', thirty days being the conventional length of the lunar month. scarcely be doubted that, as Prof. Winckler thinks, Four and Seven represent different aspects of the Moon-god, the former the four phases of the moon, the latter the seven-day week as a lunar quarter. If it be questioned whether the worship of Sibitti extended to the West, then it may be remembered that, in the list of kings of the West whom Tiglath-Pileser III mentions as paying tribute, the king of Gebal bears the name of Sibittibi'li, i. e. 'Number Seven is lord'.3

Surely, then, we are justified in explaining Bath-Sheba as 'daughter of Number Seven', Eli-Sheba', 'God is Number Seven', and, most important and interesting of all, Jeho-Sheba', 'Yahwe is Number Seven'. If this be so, we have definite corroboration of a conclusion to which many indications seem to point, viz. that Yahwe was in origin identical with the Moon-god.

I may claim, in the evidence here brought together, confirmation of two inferences I drew two years ago in my article entitled 'A Theory of the Development of Israelite Religion in Early Times'

1 Thureau-Dangin Lettres et Contrats de l'époque de la première dynastie Babylonienne No. 15. Cf. further references in Jastrow Die Religion Babyloniens und Assyriens i p. 173.

(J. T. S. ix pp. 321 ff); firstly that the Yahwe of Abraham was originally connected with the deity Sin', and secondly (as witnessed by the antiquity of the names Kiriath 'Arba, Beer Sheba, and their association with the Patriarchs), that this Deity was known and worshipped in Canaan prior to the settlement there of the tribes of Israel. C. F. BURNEY.

ASTRONOMICAL EVIDENCE FOR THE DATE OF

THE CRUCIFIXION.

IN the Journal of Philology xxix (1903) pp. 100–118, I discussed the date of the crucifixion from the point of view of technical and astronomical chronology. A discussion of the same question, partly based on my article, was contributed by Dr Bacon under the title of Lucan versus Johannine Chronology to the Expositor, Seventh Series iii (1907) pp. 206-220. In both articles it is maintained that the beginning of each Jewish month was determined empirically, and both articles depend on calculations, made by me, of the first appearance of the moon in every month which can possibly be regarded as the month of the crucifixion. In my article I expressed regret that there was no table in existence, shewing the depression of the sun below the horizon at moonset, or the altitude of the moon above the horizon at sunset, necessary to render the moon visible to the naked eye, and, in the absence of exact data, I fell back upon a vague rule given by Hevelius. Since then I have made an investigation of seventy observations of the visibility or invisibility of the young moon, made for the most part by Julius Schmidt at Athens and published in August Mommsen's Chronologie (1883) pp. 69-80. My discussion of these observations will be found in Monthly Notices of the Royal Astronomical Society lxx (1910) pp. 527-531. In this paper I found that the conditions of visibility may be expressed in terms of the difference in true azimuth and true altitude of the sun and moon at sunset, and I tabulated my conclusion as follows:

This solution is independent of differences in latitude, but not necessarily of differences in the clearness of the air between one place and another. The striking uniformity of the Athenian observations suggested, however, that the problem is almost purely astronomical, and not atmospheric. Happily, the same problem engaged the attention of Maimonides,1 who, though silent as to the observations on which his conclusions rest, gives a detailed rule for determining the date of the first visibility of the moon in Palestine. The result, according to his theory, depends partly on the true elongation or difference in true longitude between sun and moon, and partly on the apparent angle of vision at the moment when the moon might be expected to appear. This moment is, according to him, on an average twenty minutes after sunset. By the angle of vision he appears to mean the apparent difference in zenith distance between sun and moon.

Maimonides's conclusion may be summarized as follows:-If the angle of vision exceeds 11°, the moon is visible; if the angle of vision is between 10° and 11°, the moon is only visible if the elongation exceeds 12°; if the angle of vision is between 9° and 10°, the moon is only visible if the elongation exceeds 13°; if the angle of vision is less than 9°, the moon is only visible if the elongation exceeds 24°.

Converting this rule to the form in which I have expressed mine, I get :

There is no reason for doubting that the rule given by Maimonides is the result of trustworthy observations. By making a leap of a degree at a time, it gives a somewhat discontinuous result, and therefore cannot be pressed in detail, but it appears to shew that the conditions of observation are slightly more favourable at Jerusalem than at Athens. Maimonides also gives rules for computing the moon's elongation and angle of vision, but they are very inaccurate compared with the methods of modern astronomy.

1 In his treatise on the Sanctification of the New Moon, translated into German, with astronomical comments, by von Littrow in Sitzungsberichte der Wiener Akademie, Math.-Naturw. Classe, lxvi (1872) Abth. ii pp. 459-480.

The following table shews the true altitude of the moon at sunset and the true difference in azimuth of sun and moon at sunset at Jerusalem for the first two sunsets after the new moon of Nisan in each of the years, 26-35 A.D. In the case of 26 and 29 I have computed for two different new moons, as there may be some doubt as to the identity of the new moon of Nisan in those years. The table also shews on which evenings the moon ought to be visible according to my formula, and on which days of the year and week Nisan 14 ought to fall.

There is a certain amount of dispute among astronomers about the correct values for some of the constants used in the above computation. I have endeavoured to select the best constants in each case, but no set of constants would give altitudes differing by so much as o°.2 from mine, and the correction to the difference of azimuth would be much more minute.

It will be observed that in each instance the moon lies well on one side or other of the dividing line between 'Visible' and 'Invisible', as given in my summary table; there can, therefore, be little doubt that, except where the first appearance was delayed by clouds, it took place on the day specified above. On 27 March 27 the moon would, it is true, lie just on the dividing line which I have deduced from Maimonides's rule, although she stands one degree below the line as resulting from the Athenian observations. But, as suggested above, the line deduced from Maimonides appears to require smoothing, in which case the moon would be invisible on that date according to his rule as well as according to that deduced from the Athenian observations.

A comparison of the above table with the figures that I published tentatively in the Journal of Philology will shew that in six instances, viz. 28, 29 twice, 30, 31, and 33, I formerly placed the first appearance of the moon one day too late. On the other hand, Wurm, who was content to allow a minimum interval of thirty-six hours between new moon and first appearance, was in error in six instances only.1 Salmon,' who is followed by Mr Turner, allowed a minimum interval of thirty hours, which should have given him an erroneous date in the year 33 only. But he saves this error by quoting two alternative dates for that year. While it is clear that in my former paper I attached too much weight to causes other than the age of the moon, it is also clear that Salmon's success with a calculation based on age alone was largely a matter of luck. Thus the moon of 34 March 9 was new three hours later in the day than the moon of 29 March 4, yet it had an altitude of 10°.1 as against 5°-6 at the following sunset, and of 21°.3 as against 15°-9 at the sunset of the evening on which it would be visible for the first time. The moon of 34 March 9 would in fact be visible in the latitude of Jerusalem, though in another longitude, about sixteen hours after it was new, while the moon of 29 March 4 would take thirty hours to reach an equally favourable position. But as the appearance at Jerusalem could not take place till just after sunset, the moon of 34 would actually be thirty-six hours old and that of 29 thirty-nine hours old when first seen. In this way both appear to obey the rule that the moon becomes visible at the first sunset not less than thirty hours after

new moon.

It will be observed that if Nisan began on the evening on which the moon ought astronomically to have become visible for the first time, there is not one of the years under discussion in which Nisan 14 would fall on a Thursday, so that it would appear impossible for the crucifixion to have taken place on Nisan 15, as the synoptic gospels seem to imply. This date can only be saved, either by placing the first appearance of the moon in 27 on March 27, a date, which, as has been seen, is on the margin of possibility according to Maimonides's rule, if pressed literally, or by assuming that the moon was obstructed by clouds on 34 March 10, and that Nisan in consequence did not begin till the following evening. In order to render this possible it would be necessary to assume that the first appearance in the previous month was on February 9, for, if it were on February 8, the thirtieth day of Adar would close on the evening of March 10, and no Jewish month was permitted to contain more than thirty days. The moon appears to have narrowly failed to be visible

1 See his paper in Bengel's Archiv für die Theologie ii (1817) p. 293.

2 Introduction to the New Testament (1894) pp. 255-257

3 Dictionary of the Bible (1898) p. 411.