In the more fully developed calendar there are not seldom found periods of time which are reckoned without reference to any of the factors given by Nature. Such are, for example, our months, which, though historically arising from the lunar month, are now only periods of time with a definite number of days, independent of the moon. Such also is our shifting seven-day week, which, chiefly through the agency of Mohammedanism, has also been widely extended among peoples of a lower stage of development. These artificial periods, arising often from a natural period which for purposes of the calendar has been detached from its natural basis, belong to a highly developed stage of time-reckoning. Only among certain comparatively far-advanced, semi-primitive peoples does an artificial period of the simplest kind first appear, and then only one, the market-week, the origin of which it is very easy to understand.

The market-week appears in two widely separated districts—in West Central Africa, and in certain of the East Indian islands. Among the Bakongo the markets are four, viz. konzo, nkenge, nsona, and nkandu. These have given their names to the four days that comprise the Congo week. All the markets held on a certain day all over the Lower Congo are called konzo, all on the next day nkenge, etc. These markets are held at different places, e. g. all the konzo markets are held on different sites from all the markets held on the three successive days, and are so arranged that one in four will be within two or three miles of a town, the next day’s market may be ten miles away from the first town, but near some other town or towns, the next from 15 to 20 miles, the next perhaps 25 miles away from the first town. Thus every village has at least one market during the week within a reasonable distance of its doors. In order to describe the markets the place-names are sometimes added, e. g. nsona Ngungu. Each market has its special wares[1092]. The Babwende have the same names[1093]. Three Bantu tribes of the Congo State have the four-day week, but in certain cases with different names; one of the days is market-day[1094]. This is a very practical arrangement, which must gradually have regulated itself. There are also greater markets which are held every eight days[1095]—a doubling of the period, therefore. The same is the case among the Edo-speaking peoples, among whom the week is everywhere a recognised period of time, and is, properly speaking, 4 days long, this being the interval between the two markets at any given spot. Occasionally, as in the Ida district, eight-day markets are found, but the names applied to the intervening days clearly shew that a four-day week was the primary one. One of the four days is commonly known as the rest-day, and on this day men frequently stop at home, though farm-work is not absolutely forbidden. Women, on the other hand, go to market as usual[1096]. Among the Ibo-speaking peoples the names of the four days are eke, oye, afo, and nkwo. These are the same names as those of the Bini, but afo and oye are in the inverted order; it is idle to speculate on the origin of the names[1097]. In Loango the four days are variously named, but principally they are called nssona, nduka, ntono, nsilu, which names are also often applied to the open spaces where markets are held on the days in question; nssona corresponds to our Sunday[1098], i. e. it is a day of rest.

The Yoruba have, besides the market-week, a longer one of 16 (or 17) days. Of these two periods Ellis says:—The Yoruba week consists of five days, and six of them are supposed to make a lunar month, which however always begins with the new moon. (This is therefore the familiar round number.) The days are:—1, ako-ojo, the first day, day of general rest, considered unlucky; the temples are swept and water is brought in procession for the use of the gods. No business of importance is ever undertaken on this day. 2, ojo-awo, ‘day of the secret’, sacred to Ifa. 3, ojo-Ogun, 4, ojo-Shango, 5, ojo-Obatula, i. e. the name of a god, added to the word ‘day’. Each of these four days is a day of rest for the followers of the god to which it is dedicated, and for them only, but ako-ojo is a day of rest for all. Markets are held every fifth day in different townships, but never on the ako-ojo. From this custom has arisen another mode of computing time, namely by periods of 17 days, called eta-di-ogun (‘three less than twenty’). This is the outcome of the Esu societies, the members of which meet every fifth market-day. The first and fifth market-days are counted in, and thus the number 17 is obtained. For instance, supposing the second day of a month to be a market-day, the second market would fall on the 6th, the third on the 10th, the fourth on the 14th, and the fifth on the 18th. The fifth market-day, on which the members meet, is counted again as the first of the next series. These clubs are so common that the 17-day period has become a kind of auxiliary measure of time[1099]. The account contains an inward contradiction. Ellis enumerates five days and says that the market is held every fifth day, but when he reckons the days again below, the periods are four-day periods. We must probably assume that the word ako-ojo is applied to one of the four days, denoting it to be a day of rest, and that Ellis, when he says that the market is held every fifth day, is counting inclusively according to the linguistic usage of the natives, as the Greeks also did. This is the opinion of another authority, who writes as follows:—Some say the Yoruba week is composed of four days, and some of five. This same mystification recurs in the number of days said to complete one of their months. Some say there are sixteen and others seventeen days in a month. The natives rest on the fifth day, that is to say, having counted four days, they really rest on the first day of the next week, counting that day as one. So in their next great division of time they say that they rest on the seventeenth day, which is a great market-day, and this is, of course, the first day of what is their second so-called month. Fourteen of these months complete the ancient Yoruba so-called year of 224 days[1100].

But there are also periods of time of other durations. The Adeli of the hinterland of Togo divide the lunar month into five weeks of six days[1101]; unfortunately the brief account tells us nothing of the nature of this six-day week. The Tshi-speaking peoples usually reckon time in periods of 40 or 42 days, every fortieth or forty-second day being a festival termed the great adae, 18 or 20 days after which is the little adae. The great adae is always celebrated on a Sunday, and the little adae on a Wednesday[1102]. Once again the statements are not clear. If the last condition must be absolutely fulfilled, the period of the great adae must always embrace 42 days and the little adae must fall 18 days after it. The natives consider the number 40 particularly lucky and always endeavour to connect it with some important event[1103]. The probable explanation is that 40 is used as a round number instead of 42. But among the Edo-speaking peoples also, at one point in Northern Nigeria, a twenty-day month seems to be used[1104]. The former mode of reckoning is connected with the seven-day week adopted by the Tshi-speaking peoples, though this, in order that it may cover the lunar month, is reckoned in a curious fashion so that each week consists of 7 days 9 hours; each so-called day is therefore somewhat longer than the natural day and consequently also begins at a different hour of the natural day. Hence the two adae also begin at different hours of the day. The same curious reckoning is found among the Gã-tribes. This mode of computation is a far from primitive refinement, the real object of which is the fitting of the seven-day week into the lunar month, the natural day however being abandoned. There is connected with it a strong day-superstition. The first day of the ‘week’ is rest-day, and that on which the new moon falls is an absolute rest-day, the following being days of rest only for certain trades, e. g. the second for the fishermen, the third for the agriculturalists[1105]. It is clear that the only period which can pass as native is the four-day market-week, with its development the 16-day period, and perhaps also the too little known 6-day week.

In Java, Bali, and Sumatra there is a five-day market-week called pasar, in Bali also a four-day tjaturwara[1106]; alongside of these the seven-day week is in use. But wherever among heathen tribes a ‘week’ is spoken of, this is always the market-week[1107]. In Java and Bali the pasar-week is combined with the 7-day week in divisions of 35 days. Six of these periods form a wuku, a kind of year of 210 days. Besides these there are still other divisions, which are of importance for the sooth-sayers. The non-Islamite Lampong of Sumatra combine the pasar-week with the lunar month, which is counted as 30 days[1108]. We have here nothing to do with the highly developed time-reckoning of those peoples that drew up their systems under Indian and Islamite influence. This five-day week has a very extensive use in Further India: we meet it in Tonkin, in the Lao states of northern Siam, in Upper Burma among the Shan; further in Celebes and in certain parts of New Guinea. In the Malay Peninsula there is a five-day period for the determination of lucky and unlucky days. In other parts of New Guinea and in the Gazelle Peninsula of New Pommern the market takes place every third day. Of market-days in Polynesia there are unfortunately only uncertain accounts[1109].

In ancient Mexico a market was held every fifth day at every important place, just as in Africa on different days in neighbouring districts; the day was a rest-day, and with the market games and amusements were associated. This five-day market-week appears also in other parts of Central America. The Muysca of Bogota in Columbia, on the other hand, held markets every third, and the Inca peoples every tenth, day, when the country-folk ceased from labour, assembled in the towns, and engaged in traffic and games[1110]. These three- and ten-day periods are said to be brought into connexion with the month; if this statement be correct, they are not continuous periods, and the market-day must sometimes have been pushed out of place in order to secure the agreement with the moon; but the certainty cannot be ascertained.

The market-week exists therefore, as we should expect, only among peoples with a more fully developed commerce and trade. The rule attains greater importance for the time-reckoning only when, as in the East Indies, it is introduced into an already existing calendarial system. In Africa larger divisions of time have arisen on the basis of it, and in one case, that of the Yoruba, the agricultural year has been thus divided. The market-weeks, however, may also occur independently, alongside of the calendar, like the Roman nundinae, which were held every eighth day and took their name (from novem) from the inclusive reckoning.

The question of the Israelitish sabbath is complicated and has been much discussed as a point of connexion with the Babylonian civilisation. In Babylonia one day in the month was called shabattu, and the seventh day was specially distinguished. The statement that there the seven-day week existed, but as a fixed subdivision of the month, is often heard, but is an invention. I borrow the material from Landsberger’s section on the month in religious worship. A cylinder of Gudea already mentions a festival of the opening of the month in Lagash, festivals in honour of the goddesses Bau and Nina are celebrated in special new-moon houses. At all times, and later too, the day of the new moon is a great festival-day. At the time of the dynasty of Ur, under the empire of Khammurabi, and later, sacrifices were offered on the fifteenth day, the day of full moon. This is called shabattu, which word in the time of Assurbani-pal also denotes the full-moon day without any religious implication. We also find at the time of the dynasty of Ur occasional sacrifices on the day of the ‘going to sleep’, i. e. of the disappearance of the moon. These are the three days marked out by the great phases of the moon. According to them the month is divided into two halves. A Babylonian peculiarity is that the seventh day of the month, as at the time of the dynasty of Ur and under the empire of Khammurabi, becomes a day of special sacrifices. It is called sibutu, ‘the seventh’, cp. Assyrian sibittu, ‘seven’ (fem.). The 1st, the 7th, and the 28th are therefore of religious importance; for a similar emphasising of the 21st testimony is as yet lacking; instead of the 14th we have the 15th. Later, after ancient Babylonian times, the 7th becomes a day of taboo, the number 7 is made an unlucky number, and the schematic series 1, 7, 14, 21, 28, and 19 of the following month is formed (30 + 19 = 49 = 7 × 7). Hence the 14th is also sometimes designated as the day of full moon. Thus, for example, in the Creation epic, tablet 5, vv. 12 ff.:—“At the beginning of the month shine in the land. Beam with thy horns, to make known six days. On the seventh day halve thy disc. On the fourteenth day thou shalt reach the half of the monthly (growth);” in what follows the indications of the days are unfortunately missing. It is clear that the septenary division has not arisen from the phases of the moon, but on the contrary the phases of the moon have been arranged in accordance with the septenary scheme. They might also be arranged according to a quintuple scheme. Thus the tablet III R 55, no. 3[1111]:—“Sin at his appearance from the first to the fifth day, five days, is crescent,—Anu; from the sixth to the tenth day, five days, he is kidney,—Ea; from the eleventh to the fifteenth, five days, he covers himself with the shining royal cap.” It is significant of the phases of the moon that have arisen on genuinely primitive grounds that, since they are originally concrete, they do not divide themselves into symmetrical groups of days. Here the numerical scheme has been at work, and this cannot be referred to the phases, since these give no other naturally grounded divisions than the halves of the month.

The derivation of the Israelitish sabbath from Babylonia therefore offers two difficulties:—1, in regard to the word, Babylonian shabattu means the day of full moon, in fact the fifteenth day of the lunar month, and Hebrew shabbat, so far as we know, the seventh day of a period that is shifting in relation to the lunar month; 2, in regard to the period of time, in Babylonia the septenary scheme is a fixed division of the lunar month; among the Israelites it is, so far as we know, shifting, continuous, and independent of the lunar month.

I have emphasised the phrase ‘so far as we know’ since in reality our sole knowledge in this direction of the Israelitish times before the Exile is that a festival and rest-day called the sabbath existed: of its nature we know nothing. The earliest evidence we have of it is the story of one of the miracles of Elisha[1112], from which it appears that the adherents of the prophet were accustomed to gather round him on this day and at new moon, doubtless since both were rest-days. In the same way sabbath and new moon are mentioned together as festival days in Amos VIII, 5, Hosea II, 11, Isaiah I, 13. The writers during and after the Exile are the first to mention the sabbath as the seventh day of a continuous seven-day week. It has at that time the character of an ascetic rest-day, where the rest is not a joy but a duty.

Any further advance can only be made by way of hypothesis. Thus the sabbath of the times before the Exile was either, as later, the last day of a seven-day period that was shifting in relation to the lunar month, or else it was something different. Both statements are hypotheses. And if it was something different we are driven to a still further hypothesis in order to decide what it was. The suggestion most in favour is that it was the day of full moon. The sabbath is said to be the second principal day of the course of the moon simply because sabbath and new moon are always mentioned together in the days before the Exile. But this obviously proves nothing. It has further been stated that the sabbath must be a fixed day of the lunar month, since otherwise it would sometimes coincide with the day of new moon; but evidently the expression ‘new moon and sabbath’, however formally interpreted, does not in itself exclude such a coincidence. Further sabbath and shabattu are the same word, and consequently a second hypothesis is that ‘sabbath’ as well as shabattu means the day of full moon. The proof is only binding if the word in itself must mean ‘full moon’; the etymology however is disputed, so that it gives no help. It is not difficult to establish a general fundamental sense which will fit in both with the festival-day of full moon and of the seven-day period.

On the ground of the researches here carried out, however, we may put a question a satisfactory answer to which is demanded by the hypothesis just mentioned:—How is it possible for a period which forms a fixed subdivision of the lunar month to become detached from the moon and be made into an independent period shifting in relation to the lunar month? And there will still be a preliminary question to get rid of, viz. how has the septenary period arisen from the day of full moon, the 15th day of the month? The answer will be, I suppose, that the 14th, not the 15th, was taken as the day of full moon and that Babylonian influence introduced the septenary division, so that the name of one of the septenary days, the 14th, has been carried over to the rest. But since in the legislation of the Exile the great festivals were appointed for the 15th, it is clear that this day, and not the 14th, was at that time taken as the day of full moon. The question whether any late Babylonian speculation in numbers may have exercised a determinative influence upon the Jewish legislation must be decided by experts. From the unsatisfactory answer to the preliminary question I return to the main question. A shifting reckoning of this kind can only be understood chronologically as a breaking away from the concrete phenomena of Nature, an incomplete calculation being established instead of the empirical observation, as was the case, for instance, with the Egyptian shifting year, put in place of the solar year, and bringing with it months of thirty days in the place of lunar months. Now the Israelites have always had the lunar month. That a day determined by the moon should be detached from the living lunar month and made into a shifting seven-day week is quite incomprehensible and entirely without analogy. The Babylonian septenary days do not help us here, since they always remained days of the lunar month. In the light of the foregoing investigations into primitive chronology such a process would be a sheer miracle.

It remains therefore to regard the creation of the seven-day week as an act of pure volition on the part of the makers of the refined exilian legislation, who took the name of the ancient sabbath, a festival-day of uncertain position, and applied it to the seventh day of a shifting period. And this is equally difficult either to prove or disprove. It is seldom found that a new creation proceeds entirely from nothing, and no analogy to the shifting seven-day period is anywhere to be met with—except in one case to be mentioned presently, the market-week. Especially in matters chronological would it appear that the Jewish legislation did not radically break with antiquity, but systematised and cultivated already existing tendencies, if we may judge by the few points of departure handed down from the earlier period; hence the numbered months, hence the fixing of the great festivals on the day of full moon. We are speaking here not of the changed religious character of the sabbath, but of the chronological question. If therefore fundamental grounds are lacking for the creation of a shifting seven-day period by the legislation of the Exile, we must cling to the other hypothesis, viz. that in pre-exilian times also the sabbath was the seventh day of a shifting period, which the legislation has transformed in its own fashion.

But if the shifting sabbath is old, the question arises whether analogous periods exist in primitive time-reckoning. Certainly they do, and they are periods of a quite definite nature,—the market-weeks. There are market-weeks of three, four, five, six, eight, and ten days: that seven does not appear in any example must therefore be an accident. The market-week is spread over the whole earth at a more advanced stage of civilisation. The market-day is a rest-day, since the people go to the market: since they rest and gather together it is therefore a festival day. So also with the Roman nundinae, on which no public meetings were held and the schools were closed. The dispute of Roman scholars as to whether the nundinae were religious festival-days or business-days is significant[1113]. Since the market-day is a day of rest, however, it is also, as in West Africa, made a taboo day on which work is forbidden. The connexion between the market and religion is universal and appears particularly clearly in heathen Arabia[1114]. It is true that no market-day is attested for ancient Canaan, but even in pre-Israelitish times the land was already covered with towns, so that the conditions for regular markets were the same as in ancient Greece and Rome. From post-Biblical times at least three great annual markets are known; one was held at the terebinth of Hebron, which was at the same time the object of a cult. In Midrash it is allowed to visit a heathen yearly market at the half-holidays of the Passover and of the feast of Tabernacles[1115]. Since the day was a rest-day, the command for rest might gradually, through a new interpretation, be applied to the original purpose of the market, viz. trade. In Amos VIII, 5 the traders complain:—“When will the new moon be gone, that we may sell corn? And the sabbath, that we may set forth wheat? making the ephah small,” etc., but the command for the absolute sabbath’s rest was certainly not carried out at that time, nor yet in the time of Jeremiah[1116]; after the overthrow of the Jewish monarchy the trade of the markets on the sabbath revived, if indeed it had ever perished. Nehemiah, three centuries after Amos, has to give the injunction:—“ ... and if the peoples of the land bring ware or any victuals on the sabbath day to sell, that we would not buy of them on the sabbath, or on a holy day[1117],” and the breach of this law is sternly reprimanded:—“In those days saw I in Judah some treading wine-presses on the sabbath, and bringing in sheaves, and lading asses therewith; as also wine, grapes, and figs, and all manner of burdens, which they brought into Jerusalem on the sabbath day.... There dwelt men of Tyre also therein, which brought in fish, and all manner of ware, and sold on the sabbath unto the children of Judah, and in Jerusalem.” Nehemiah reproves the nobles:—“Did not your fathers thus, and did not our God bring all this evil upon us, and upon this city?”, and he has the gates shut and guarded when it grows dark before the sabbath. When, notwithstanding this, the merchants once or twice encamped outside the walls on the sabbath, he drove them away with threats[1118]. At this time work was performed and trade carried on on the sabbath, though certainly it does not follow that the sabbath was the principal market-day of the week: we are speaking of a large town, where no doubt there was a market every day. But it would be quite in keeping if in smaller matters the sabbath had once been the proper market-day.

The work of Webster culminates in an attempt to explain the sabbath. The author brings together abundant material for the practice of assigning certain taboos to certain days, partly notable days in the experience of human life, such as birth, death, etc., and partly those regularly recurring days which are dependent on superstitious and religious ideas. Among these days are found both the market-day and the days of the principal phases of the moon,—the day of new moon, in a lesser degree the day of full moon, and further also the days of the darkness, of the moon’s invisibility. He rightly distinguishes the continuous Israelitish week from the ‘unlucky days’ of the Babylonians, but is nevertheless of the opinion that the sabbath is really the day of full moon, which in this character was overlaid with certain taboos and has become independent of the moon. How this separation was effected, Webster does not explain: he merely makes the statement. He has not felt the decisive difficulty, which lies just in this point, because he has not attacked the problem from its chronological side. There is no reason to suppose that the day of full moon could become detached from the genuine lunar month, and such a process would seem still more strange since the day of new moon remained a genuine new-moon day. On the other hand the development of market and rest-day into a day of taboo is everywhere natural, and is attested in the above examples from Africa; this taboo character was emphasised and inculcated by the late Jewish and exilian legislation in opposition to the old festive merry-making. The new-moon day, which had fallen out of the scheme, was at the same time rejected and proscribed. The suggestion that the sabbath arose from the market-day is certainly only a hypothesis, since a definite market-day is not demonstrated for Canaan; but it has the advantage of remaining within the limits of primitive time-reckoning, which knows no other continuous periods than the market-weeks.

Festivals and time-reckoning are from the beginning inseparably bound together. Some of the former have already been dealt with, e. g. the festivals of the new moon, the full moon, and the beginning and end of the year. It remains briefly to sketch the development of this connexion and to illustrate it with a few examples. A detailed discussion would lead us too far away from the main theme into the domain of the history of religion. How many pages have been written about the New Year festival alone!

The connexion between festivals and time-reckoning is grounded in the fact that both are originally dependent on the phases of Nature. Festivals are already held at definite times of the year by peoples who know nothing of a proper time-reckoning, e. g. the much-discussed Intichiuma ceremonies of the aborigines of Australia. They are closely associated with the breeding of the animals and the flowering of the plants with which each totem is respectively identified, and as the object of the ceremony is to increase the number of the totemic animal or plant, it is most naturally held at a certain season. In Central Australia the seasons are limited, so far as the breeding of animals and the flowering of plants is concerned, to two—a dry one of uncertain and often great length, and a rainy one of short duration and often irregular occurrence. The latter is followed by an increase in animal life and exuberance of plant growth. In the case of many of the totems it is just when there is promise of approach of the good season that it is customary to hold the ceremony. The exact time is fixed by the alatunja (the chief of the local group)[1119]. The ripening of a plant which is an important article of food is often accompanied by certain ceremonies by which the eating of the fruit is first made lawful. These so-called sacrifices of the first-fruits, which have been touched upon above[1120], are therefore dependent upon a definite natural phase, and there may be several of them in the course of the year.

At seed-time a festival is celebrated in order to secure the good growth of the seed. The Bahau of Borneo, who have the agricultural year[1121], celebrate two great festivals, one at the sowing (tugal, from nugal, ‘to sow’), and one after harvest, the festival of the new rice-year, dangei, which however is not held if the harvest has failed; it is the climax of the year. At both festivals the people gorge themselves to the full, rice being given even to the animals. But during the period of growth also the plants need protection and blessing, various plants require and obtain different festivals, so that a cycle of agricultural festivals arises[1122]. The southern tribes of the Malay Peninsula celebrate three great agricultural festivals in the year, one after the transplanting of the young rice-plants, another after the formation of the fruit, and a third after the harvest[1123]. As an example of a fully developed festival-cycle of this kind I give the festivals of the Bontoc Igorot, with which should be compared the section on the agricultural year of this tribe[1124]. After the conclusion of the time when rice-seed is put in the germinating beds, pa-chog, the festival po-chang is held, after the transplanting of the rice the festival chaka (held on Feb. 10 in 1903), and after that an unexplained festival su-wat; on the day on which the first ‘fruit-heads’ have shown themselves on the growing rice there is the festival ke-eng, and on the following day tot-o-lod; sa-fo-sab, before the beginning of harvest, introduces the harvest. At the end of the rice-harvest and the beginning of the period called li-pas (‘no more rice-harvest’) lislis is celebrated; at the time of the planting of camotes loskod; in the same division of the year, called bali-ling, the festival o-ki-ad, when black beans are planted. Finally at the end of this division we have ko-pus, a three day’s rest, just before the work of rice-culture is begun again[1125]. An African example from the neighbourhood of the Lower Niger will shew how in this agrarian festival-cycle other feasts arise which may in part be older. The cycle consists of the following festivals:—1, sacrifices and adoration to the great spirit or creator, always made in anticipation of the new crop, to ensure that it is good; 2, communion of first-fruits, a festival to the house-hold gods; 3, communion of the new yam; 4, the feast of hunters; 5, ofala, a celebration to Ofo, god of justice and right, in honour of the public appearance of the king; 6, the crumbo, or remnants of yam, reserved for the king only; 7, the feast of roast yam at the close of the year, the termination of this marking the end of the native year and the feast also serving as a form of public notice that farming has to recommence. This is a festival in honour of Ifejioku, god of the crops, as a token of gratitude on the part of the community for a fruitful and prosperous year. It is usual for the king to give a month’s notice before each ceremony takes place[1126].

A pastoral people may also have a well-developed festival-cycle marking the points of the year which are important for their herds. I quote as an example the main festivals of the Reindeer Koryak of Eastern Siberia. There is a ceremony on the Return of the Herd from the summer pastures, when the first snow covers the ground. In spring, when the fawning period is over and the reindeer have lost their antlers, the fawn festival is celebrated. The fire in the house is put out and a new one started by means of the sacred fire-board. Some tribes pile up the antlers of the slaughtered reindeer. Other festivals are observed:—1, when the sun marks the approach of summer after the winter solstice: a sacrifice is then offered to the sun; 2, in the month of March, when the does commence to fawn: a sacrifice is offered to The-One-on-High; 3, in spring, when the grass commences to sprout and the leaves appear on the trees; 4, when mosquitoes put in their appearance—reindeer are then slain as an offering to The-One-on-High, lest the mosquitoes scatter the herd[1127].

Here the development is simple and clear, but not so among many peoples where agriculture or the raising of cattle does not occupy so important a place. The Maidu of northern California have four seasons and four festivals founded by the hero Oankoitupeh:—‘the open air festival’ in the spring, ‘the dry season festival’ about the first of July, ‘the burning to the dead’ about the first of September[1128], and ‘the winter festival’ about the last of December[1129]. The connexion with the seasons is clear, but we do not even know whether the names are of genuine native origin. This example clearly shews that the great difficulty lies in the fact that the real nature of the festivals is unknown. But often where detailed accounts of a festival exist, the original reason for it becomes obscured in the course of the development, so that the original connexion between festival and season cannot be established. This is especially the case with peoples among whom the religious life has had an especially strong development.

A phenomenon peculiar to the peoples of the far North is that the winter is the time of the festivals. The summer is the good season, when supplies for the winter must be collected; it is therefore a very busy time, when each family has to work for itself and has no leisure for festivals. The winter is the time of rest, in which the people live on the supplies already collected; they naturally crowd closer together, and have much leisure, which is used for religious ceremonies and for games. Hence the winter is the time of the religious ceremonies among the Eskimos, the Tlinkit, and other Indians of N. W. America[1130], and hence the Yule festival celebrated in the winter becomes the greatest festival of the Scandinavian peoples[1131].

When a festival takes place, people assemble together who often have to come long distances. We have spoken above[1132] of the devices adopted in order to ensure that the day of an appointed non-periodic festival shall not be missed. Periodically recurring festivals, which are connected with a natural phase or some occupation, particularly if this is agricultural, are determined as to time, but not accurately. Hence it is already found among the Central Australians that the exact day is fixed by the chief. Such festivals, appointed within certain limits assigned by Nature, are found also among peoples with a fixed calendar, e. g. the Roman feriae conceptivae. Significantly enough, these are agricultural festivals which, on account of the change of position of the lunisolar year in relation to the natural year, could not well be regulated by the former. But where a calendar exists, this is the given means of regulating the festival dates so that preparations can be made and the people can assemble at the right time. In the natural and agricultural years the festivals are in the proper sense conceptivae; the question is properly to find a means of accurately fixing the day within the short periods given by Nature. This purpose is served by the calculation from the moon. The moon herself has her festivals, especially that of the new moon and, though more seldom, that of the full moon[1133]. Thus the festival times are regulated by the moon. In itself any suitable day of the month can be appointed as a feast-day, but custom and superstition cause certain days to be preferred. Thus the day of new moon, since it was often already a feast-day in itself, was bound to be preferred. The Natchez of Louisiana, for instance, celebrated at each day of new moon a feast which took its name from the animals and plants which the preceding month had principally brought forth, but the greatest festival was that held at the new moon of the first month.[1134]

It is a very wide-spread idea that things which are to prosper and grow should be undertaken during the time of the waxing moon, and that anything begun when the moon is on the wane will dwindle and die. Hence the proper time for a festival is the bright half of the moon, and especially the time at which the moon has attained her full shape. It is not only on account of the fair light which costs them nothing that the negroes dance on the nights of full moon. In Dahomey the festivals take place at the full of the moon, and the days are determined by the native government[1135]. In Burma all religious festivals with the exception of the New Year festival, the date of which is regulated in a special manner, take place at the time of full moon[1136]. Throughout Australia, Tasmania, and Melanesia the festivals begin either at full or new moon[1137].

In regard to the Israelitish festivals, the antiquity and great importance of the new moon festival has already been pointed out[1138]. The Jews here follow a wide-spread custom. Whether they, like many other peoples, also preferred the time of full moon for their festivals, is a more difficult question. A fixed day for the Passover and Feast of Unleavened Bread and for the Feast of Tabernacles is first prescribed during and after the Exile, the last-named on the fifteenth day of the seventh month, the Feast of Unleavened Bread on the fifteenth day of the first month, and the Passover on the evening of the day before (the fourteenth of the first month)[1139]. The only other information we have from ancient times as to the date of the Feast of Tabernacles is contained in the earlier name ‘Feast of Vintage’; it was celebrated after the conclusion of the fruit-harvest and vintage. In regard to the Feast of Unleavened Bread—since it is with this chiefly that we have to do, not with the preliminary Feast of the Passover associated with it, which was a feast of a different nature—the order of the Yahwist runs ‘at the time appointed in the month Abib’[1140]; as a motive is adduced the fact that the Jews came out from Egypt in this month. The Deuteronomist[1141] transfers this to the preliminary festival. The time therefore, like that of the Feast of Vintage, is determined by an event in agriculture, but at the same time by the moon. Linguistically chodesh can here mean ‘new moon’; in that case we could also translate ‘at the time appointed after the new moon of Abib’; but since the sense ‘month’ is so old and the original sense ‘new moon’ appears unequivocally only where monthly new moon festivals are in question[1142], it seems reasonable to translate the word here simply by ‘month’. Now it is often stated that the festive seasons both of the Unleavened Bread and of the Feast of Vintage were regulated purely by natural circumstances: the former was celebrated when the first ears ripened, and the latter when the fruit-harvest was at an end, each according to local conditions. But the Feast of Vintage at least was a general festival even in Canaanitish days[1143], and moed properly means ‘determined, appointed time’. It was therefore not accidental circumstances but a rule that in early times called the people together to the festival. Chronological regulation is proved by the name of the festival of harvest (chag haq-qazir), ‘Feast of Weeks’, chag shabuot in the Yahwist[1144]. The regulation by the weeks, however, is late and artificial in comparison with that by the moon.

Now if we know what part was played by the time of full moon in the festivals of other peoples, and indeed for the agrarian peoples also, in spite of the differences in date resulting from the observation of the time of full moon, it seems always probable that the regulation of post-exilian times for the fifteenth originated in an old tradition in accordance with which the time of full moon was specially favoured for the feast. Earlier the date was not so accurately observed; the time of full moon was prescribed so that those who were prevented from celebrating the Feast of the Passover at the proper time might do so on the fourteenth of the following month[1145]. Unfortunately the date of the passage in I Kings (XII, 32), according to which Jeroboam celebrated the Feast of Tabernacles on the 15th day of the eighth month, is doubtful; if the passage is old, it affords valuable evidence that the time of full moon was the proper time for holding agrarian festivals[1146].

Among the Greeks all the ancient festivals with the exception of the feasts of Apollo, which always took place on the seventh of the month, were concentrated in the period shortly before and during full moon[1147]. The selection of days is organically connected with the lunar reckoning, and the superstition of days has arisen independently among different peoples. As an example the sacrifices of the Toba Batak of Sumatra may serve. At the felling of a tree for house-building sacrifices must be offered during the waxing moon; this is in general the favourable time, since everything undertaken then increases with the moon. The huntsman sacrifices to his god at noon-tide about the time of new moon, the fisherman at noon while the moon is waxing; before a military expedition a certain sacrifice is offered (preferably in the early morning) at the time of full moon, and another at the waxing moon[1148].

This superstition, which involves the accurate knowledge and observation of the days, and the injunction, to which great religious importance is attached, to celebrate the festivals on the proper days, lead to the result that the time-reckoning, which arose in the first place from the events and necessities of practical life, has among certain peoples passed completely under the influence of religion and has been further developed from ecclesiastical standpoints in the service of the religious cult.

There are however other ways of exactly fixing a day, viz. by observation of the stars and of the solstices and equinoxes. The former method is hardly ever used directly as a means of determining religious dates, and this fact is very significant for the practical character of the observation of the stars. No religious ideas are associated with the phases of the stars, although star-myths innumerable are related. The reason is not easy to discover. A contributory factor may be that although the observation of the stars is wide-spread, it is yet not a matter which concerns every man, and also that the stars always give only a single point of time and do not form cyclical periods within the year, though on the other hand they are intimately connected with the phases of the natural year and with agriculture. The principal reason may be conjectured to be that the reckoning of months, on account of its connexion with the popular festival seasons and with the selection of days, has been from the beginning chiefly carried out with a view to religious considerations.

It is only among certain peoples that the observation of the solstices and equinoxes plays any great part, and that consequently the religious importance of the sun is also great. But the festivals of the solstices and equinoxes, recurring at regular intervals in the course of the year, are far from being able to compare with those of the phases of the moon. It has already been mentioned that the Eskimos were able accurately to observe the winter solstice[1149]. At this time, about the 22nd of December, they held a festival to rejoice over the return of the sun and the good hunting weather. They collected together from all over the country in great parties, entertained one another in the best possible manner, and when they had gorged themselves to the full they got up to play and to dance[1150]. Certain Indian peoples have made quite a special custom of the observation of the solstices and equinoxes. Thus for instance did the Inca people, but they had lunar months also, and even the great festival of the sun in December was regulated by the days of the lunar month[1151]. The Zuñi determine the festival times by the observation of thirteen different positions of the sun on the horizon, but they have also lunar months, five of which are named from natural phases, and six from colours borrowed from certain rites[1152]. The ceremonies are therefore still distributed among the months, and the most obvious explanation is that the observation of the thirteen positions of the sun really serves to determine the thirteen months, and with them the times of the rites. The old Mexican calendar seems to have no connexion with the moon, but in Ginzel’s opinion this does not exclude the possibility of an earlier development on the basis of a relationship with the course of the moon[1153]. In any case the regulation of the festivals by the positions of the sun is a comparatively isolated separate development among certain peoples; the regulation by the moon, on the contrary, is found all over the world.

Because the calendar is principally looked upon as the concern of religion, the months appear in such close association with the festivals held in them that it is sometimes found that the relationship to the phases of Nature falls into the background. Among peoples who have no names of months, like the Greeks of the Homeric period, or among those who name only some of them, it may therefore happen that the months become named from the festivals or perhaps that such names supersede those which refer to natural phases. Thus, as has been mentioned above, six months of the Zuñi year are named from the colours of the prayer-sticks. Of the Inca months one is named from a moon festival, two from provincial festivals, and one from the great sun festival; the rest take their names from the occupations of agriculture[1154]. Of the tribes of Bolivia it is stated that their knowledge of the calendar is not according to days, but according to the principal festivals[1155]. In Africa two examples have been given[1156], those of the Hausa states and the Edo-speaking peoples. In the Babylonian calendar the names of months derived from festivals spread more and more, at the expense of names of other kinds[1157]. The phenomenon is therefore comparatively rare and is found only among peoples who have a highly developed religious cult, and even in the examples here given the process is not consistently carried out.

Consistency is found only in one case, the calendar of ancient Greece, and is all the more striking since in the hundreds of varying calendars of the town-states no names which do not refer to festivals have been with certainty demonstrated; the few calendars with numbered months are of more recent origin[1158]. The certain conclusion is that the Greek calendar was entirely regulated from the point of view of the religious cult. Where on the other hand the place of the lunisolar year is taken by another reckoning, it is found that the lunar reckoning is still used in the establishing of certain festivals, as for instance in Bali[1159], and by the Christians in the matter of Easter and the festivals depending thereon.