chronology must work with a system, and this fact often blinds the eye of the chronological student to the irregularity in the practical treatment of the calendar. Irregular intercalations of this kind are not indeed attested for Iceland, but it is evident that they must always appear of themselves in a defective calendar. The possibility of a treatment of this kind existed, since the spokesman of the laws had to proclaim publicly every year to the assembled people in the Althing notices about the calendar for the following year, among which the announcement of the intercalation held a special place. In these arguments I find myself in agreement with Beckman: I also agree with his statement as to the gradual increase in accuracy in the formation of the Icelandic week-calendar under the influence of the ecclesiastical calendar. We conclude then that the cardinal points of the Icelandic calendar, which recur throughout Scandinavia and fall about three weeks behind the equinoxes or the solstices, are not of Christian origin: the agreement with what Bilfinger terms the ‘mean Easter Thursday’ is accidental. The date is due to climatic conditions. A contributory factor may have been the circumstance that mid-winter and midsummer fall just at the places where a shortening or lengthening of the day becomes observable. [337] Småland and neighbouring provinces. Communicated by Dr. von Sydow. [338] This practice has passed into the Lapp language: _kess idja_ = week of the summer nights, _talvidja_ = the winter nights. Wiklund, pp. 16 and 20. [339] _Þá skylldi blóta i móti vetri til árs, enn at miðjum vetri blóta til gróðrar; hit þriðja at sumri, þat var sigrblót_--_Heimskringla_, Ynglingasaga, ch. 8. [340] See e. g. above, p. 70. [341] Coquilhat, p. 367. [342] Maass, p. 314. The names are those of the Arabic letters and also denote the years of an eight-year cycle, the years of which are said to be characterised by similar weather. The people are Islamite Malays. Astrology and the calendar have strongly influenced Sumatra and in particular Java; primitive modes of thought however recur under the surface. [343] Brown, p. 331. [344] Thurnwald, p. 346. [345] _Ibid._ [346] Routledge, p. 40. [347] Hale, p. 105. [348] Hastings, p. 132. [349] Swoboda, p. 22. [350] Brown, p. 331. [351] Skeat and Blagden, I, 393. [352] De Backer, p. 406. [353] Hagen, p. 154. [354] Brown, p. 347. [355] Parkinson, p. 378. [356] Cp. p. 57. [357] Above, p. 55. [358] _Loango Exp._, III: 2, 139. [359] Roscoe, _Baganda_, pp. 37 ff. [360] Id., _Bantu_, p. 72. [361] Schiefner, pp. 191 ff. [362] See above, p. 75. [363] Schiefner, pp. 198, 201 ff. [364] Wirth, p. 211. [365] Hale, pp. 106, 170. [366] Mathias G., p. 211. [367] Dennett, pp. 136 ff. [368] Adriani and Kruijt, II, 264. [369] Maass, p. 512. [370] Evans, _JRAI, 42_, p. 395. [371] Mommsen, _Röm. Chronologie_², pp. 47 ff.; bibliography in Ginzel II, 221 ff. [372] Schulze, p. 369.